Original post on fosstodon.org

Why do I continue to maintain #sysklogd on Linux when there's rsyslog and syslog-ng?

I believe what sysklogd has going for it is exactly what the competitors sacrificed: simplicity, a tiny footprint, zero dependencies beyond libc, and a config file a human can read in five minutes. On embedded […]

Reenviar logs de Linux a un servidor syslog » Proyecto A Cómo reenviar los logs generado en un equipo con Linux a un servidor de syslog. Indicamos el procedimiento para distribuciones Debian, Ubuntu, Red Hat, Rocky, Oracle, CentOS, etc. Requisitos para envi...

Reenviar logs de Linux a un servidor syslog proyectoa.com/reenviar-log...

Cómo reenviar los logs generado en un equipo con Linux a un servidor de syslog

#syslog #linux #rsyslog #log #reenvío #logs #eventos #servidorsyslog

Add DNS SRV discovery to omfwd by rgerhards · Pull Request #6328 · rsyslog/rsyslog Summary add a targetSrv parameter that resolves _syslog SRV records and feeds the results into omfwd's target pool process SRV priority/weight ordering, enforce mutual exclusivity with target,...

Looking into #syslog server implementations, trying to decide if I should go with #rsyslog or #syslog-ng. Apparently the maintainer (!) of rsyslog is a slop coder (see https://github.com/rsyslog/rsyslog/pull/6328 Might want to avoid rsyslog in the future, I definitely will.

#slop #ai #llm

Original post on mastodon.social

People still see documentation as “just text”. For rsyslog, that changed a bit over a year ago.

Docs now need to work for humans and AI tools: clear structure, stable anchors, semantic hints, and examples that code agents can verify.

This turned into real engineering work, not just writing.

I […]

Original post on mastodon.social

rsyslog will (most probably) soon speak YAML.

Not a revolution — just joining the languages the rest of the stack already uses.

Simple stuff in YAML, complex logic still in RainerScript.
And yes, you can mix both.

Think: easy setup for containers and cloud, full power for those who like […]

Original post on mastodon.social

rsyslog now mirrors its source in Europe and China, ensuring global access even when platforms slow or fragment. 🌍

Same code, same community — no borders, no lock-in.

Read more: www.rsyslog.com/keeping-rsyslog-accessib...

#rsyslog #opensource #linux #devops #foss #logging […]

🚀 Myth-buster: rsyslog isn’t just a “legacy syslogd”.
It’s a full-blown ETL engine for modern data pipelines — ingesting from files, journals, syslog, Kafka; transforming with RainerScript, mmnormalize, GeoIP, PII redaction; and delivering to Elasticsearch […]

[Original post on mastodon.social]

Original post on mastodon.social

As a side-activity, I worked on AI Agents for the rsyslog ecosystem. I had a really good time. We use Digital Ocean Gradient AI, for which #rsyslog got a sponsorship from #DO - makes testing even more relaxing :)

rainer.gerhards.net/2025/10/rsyslog-gradient... […]

DigitalOcean Renews Sponsorship – A Strong Foundation for rsyslog Innovation We are pleased to announce that **DigitalOcean** has renewed its sponsorship of the rsyslog project. The partnership between rsyslog and DigitalOcean has been in place for many years and continues to be an essential part of our technical infrastructure. DigitalOcean provides the compute resources that power both large parts of our **continuous integration (CI) pipeline** and the **package download server** used by thousands of rsyslog users worldwide. This support ensures the reliability and consistency of rsyslog releases and testing workflows, which are critical for maintaining a stable and trusted open source project. The renewed sponsorship gives us the stability we need for day-to-day operations while also providing flexibility to explore new ideas. In particular, it enables **enhanced experimentation** within the scope of our broader **AI-First strategy**. While AI-First stands on its own, DigitalOcean’s support and its **Gradient platform** allow us to test concepts more quickly and gather early feedback from real-world use. We are currently exploring new ways to integrate AI-driven assistance into rsyslog’s documentation and development ecosystem, using Gradient for selected prototypes and evaluations. These experiments complement our ongoing modernization work and help us understand how intelligent tools can make rsyslog easier to configure, extend, and maintain. We thank DigitalOcean for its long-term commitment to open source and for supporting both the **core infrastructure** and our **forward-looking research and development** efforts. As our experiments mature, we look forward to sharing updates and continuing this strong collaboration.

We got another round of infrastructure sponsorship from Digital Ocean. That will also cover AI experiments.

www.rsyslog.com/digitalocean-renews-spon...

#DOforOpenSource #rsyslog

Original post on mastodon.social

I have a project where I need to revist normalization of traditional Windows Event Log syslog messages. It is interesting how many more options we have thanks to new tech. Look at an overview, including why this is relevant for cloud-native pipelines […]

Outdated readthedocs problem solved! I am glad to tell that I finally managed to solve an issue that caused confusion for years. Someone had cloned and published the rsyslog documentation at readthedocs. Unfortunately, it was not maintained afterwards and also looked like an official rsyslog doc. That added a lot to the “rsyslog’s doc is bad and inconsistent” feel inside the community. This could now be resolved, and current, official doc is now available at readthedocs. I am very happy and glad for readthedocs staff members who helped us to finally resolve the issue. The current rsyslog documentation is finally shown at readthedocs. (Screenshot: 2025-09-18, Rainer Gerhards) ## AI exposes the Problem even more The problem worsened as AI grew. readthedocs is a well-respected doc site with a lot of trustworthiness in SEO and AI ingestion. That made the AI agents believe the doc there would be almost as good as on rsyslog.com. As it was a 2016 snapshot, this introduced a lot of inconsistency between the real state and what it claimed, opening a real big problem in regard to AI hallucination and simple AI picking the wrong format. I tried several times in the past to work with the original submitter, but to no avail. I think I also tried at least once to contact readthedocs, but I am sure I did it wrongly. At that time I did not have the insight and was simply left frustrated. With the rise of AI, I saw an even stronger need to correct the issue. Due to my past experience, I now used AI help to understand how to contact readthedocs and that told me how to do it right. That said, all worked swiftly, and the support at readthedocs was very professional and solved the issue according to their policies. I have now control over the project and it displays the current rsyslog documentation (yay!). That said, we will keep the rsyslog doc up at readthedocs as well. I think there are some folks who have become used to read it over there, and there may be other folks who simply prefer the site. As we can now keep the version current, and so it is great to have another vary capable mirror in place. ## Some Outdated Doc Problems remain We still have some minor issues, e.g. on readthedocs there is published v5 rsyslog doc, which of course is hopelessly outdated. I can understand the will to provide v5 specific info, even though I wonder if really someone still runs a system 10+ years outdated. I will try to contact it’s maintainer to check if he will accept a small update so that any page tells it is outdated and no longer relevant to current rsyslog. Note that this of course is not only a readthedocs issue – outdated doc and code snippets are all over the Internet and we need to fight this in order to streamline the user experience. Naturally, I everyone does this for good reason, and we appreciate the work people have done over the years to help rsyslog users. So this “fight” is one of decent demotion (by better doc and spider hints) as well as identifying top ranking pages and trying to help their owners to update them. Note that even my blog holds outdated snippets. We currently evaluate an automated method to tell this spiders. ## Quick Wrap-Up But I am sure that the now-correct readthedocs “rsyslog” project is a huge step forward into better user experience and AI ingestion.

With help from the readthedocs team, we could finally solve a nasty issue regarding outdated rsyslog doc. Read the good news and why exactly this is good news: rainer.gerhards.net/2025/09/outdated-readthe...

#rtd #rsyslog #ai

Rsyslog Assistant

"rsyslog assistant" #AI has been update. We better answers - try it at https://rsyslog.ai

www.rsyslog.com/improved-rsyslog-assista...

#rsyslog

Status update: omhttp, CI, backlog, and containers Time goes fast, it is Sep 10 already. Mid August I said we will do a great refactoring of `omhttp` within a week or two. Well, that did not work out as planned. We still made solid progress, but more pressing work put it on hold for a bit. Time for a small update of what is happening in rsyslog. Symbol image for “Status Update” type of postings. (Image: Rainer Gerhards via AI) ## What changed since Aug 26 * We shipped v8.25.0 on Aug 26 as planned. * Roughly 100 new commits already landed in daily stable since then! * Around 150 old issues were revisited and closed, some as implemented feature requests. * About 10 long pending PRs were merged. Short version: steady hands, lots of broom work, and a few shiny bits. ## omhttp progress The goal was a quick refactor sprint. Reality voted otherwise. I underestimated a few edges, but the main reason is that other work had higher priority. We paused new additions so we can put the already implemented pieces through a proper practice drill inside our containerization effort. If you want the background and the planned shape, it is all in the mid August post. ## CI and tooling I took a detour into CI. Newer runners and compilers, better coverage handling, and more reliable signals. This lines up with the expectations I set in late August: quick maintainer look, clear gates, no mass closes. It is not flashy, but it pays off every single day. ## librelp and GCC 15 librelp did not build cleanly with GCC 15 where C23 is now the default. I fixed that and updated its CI while I was there. Outcome: clean builds and fewer landmines for downstreams. ## Backlog, issues, and PRs We kept working through older tickets. Quite a few closed for good reasons, some turned into implemented feature requests. I am grateful that many PR authors stayed with us and came back to finish their work after months, one even after years of me being the bottleneck. Thank you. The process matches what I wrote in late August. ## A complex issue that was not I chased a “tricky” crash again. It looked like deep concurrency. It was a missing state update. The useful bit is the lesson: tighten contracts, add tiny deterministic tests, and verify assumptions before hunting ghosts. ## Guardrails, still friendly We continue to see the occasional odd PR pattern. The approach stays friendly first with light automation and clear rules. That keeps the bar welcoming and protects the project at the same time. I described the guardrails and the small welcome workflow in more detail here. ## Focus drifts to project containers Focus slowly drifts toward optimizing our project provided containers. Goals: predictable defaults, smaller delta between dev and prod, and clearer guidance. This ties directly to the `omhttp` work and CI reliability. If you have container HTTP output edge cases, please send them. Partial accept and retry behavior are especially interesting. ## Infrastructure note We are also setting up new infrastructure. Thanks to DigitalOcean for their continuous sponsorship, which makes this much easier. We even have a small helper issue to prep fresh Droplets for our test and staging work. Much appreciated. ## Docs and AI assistants Doc restructuring continues with near daily commits. Smaller pages, clearer parameter docs, and a structure that is friendly to both humans and AI helpers. We also added more AI assistants behind the scenes. I will introduce them once the dust settles. Promise to keep the jokes human. ## What is next * Finish and merge the remaining `omhttp` bits, then document good defaults and migration notes. * Land the remaining CI cleanups and keep coverage useful, not noisy. * Keep the steady cadence on old issues and PRs. * Harden and document the project containers. * Continue the doc split ups and metadata cleanups.

How is #rsyslog doing? A quick status update.

rainer.gerhards.net/2025/09/status-update-om...

Original post on mastodon.social

Looks like #rsyslog is now also being trageted by social engineering PRs. This is actually good news! It helps us gain first-hand experience and prepare countermesures right from the start. #AIFirst will for sure play a positive role in that.

The dirty and interesting details in blog posting […]

Original post on mastodon.social

rsyslog 8.2508.0 (scheduled stable) is out.

Highlights:

* Prometheus stats (impstats) + imhttp health check

* Hardened TCP server; imtcp race fixes

* omelasticsearch handles 401/403; OpenSSL PrioritizeSAN

* omhttp on OMODTX; headerless detection in pmrfc3164

* Docs now under doc/ […]

Original post on mastodon.social

Honest note: my responses to PRs/issues were sluggish, especially during the pandemic. Sorry, and thanks for sticking with rsyslog.

What changes now:

- Quick maintainer look within 3 business days

- AI review runs instantly (90%+ useful)

- Full review after CI green + AI items addressed

- […]

Original post on mastodon.social

#AI companies of course want lock-in. No other way why code agents tend to have their own instruction files albeit the AGENTS.md standard exists. That's annoying.

We at #rsyslog do not like lock-in. We try to work around it, using their files, to most extent, to point the agent to existing […]

Introducing the rsyslog commit AI assistant # We are adding a new helper to our Responsible AI First toolbox: the **rsyslog commit (message) assistant**. It is an **optional** ChatGPT Custom GPT that helps contributors write **clear, policy-compliant commit messages** faster. You stay in full control. Nothing in your workflow changes unless you want it to. It follows our responsible “AI First” strategy and is optional, transparent, and keeps the human in full control. * Use it in your browser: **https://www.rsyslog.com/tool_rsyslog-commit-assistant** * Review its exact behavior (full transparency): **https://github.com/rsyslog/rsyslog/blob/main/ai/rsyslog_commit_assistant/base_prompt.txt** * Discuss and give feedback: **https://github.com/rsyslog/rsyslog/discussions** ## Why we built it * **Clarity and consistency:** higher-signal commit history and smoother reviews. * **Lower barrier for newcomers:** fewer style nits, more focus on substance. * **Good AI use:** opt-in, scoped, transparent, and aligned with our contribution rules. * **OSS-friendly cost model:** a Custom GPT lets us support contributors on a lean budget. ## What it does The assistant asks for the essentials and proposes a commit message you can edit: * **Inputs it expects:** a diff (pasted or link), optional issue URL, an optional non-technical reason, and optional context like component or whether a bug was seen in practice. * **What it produces:** a commit message that follows our rules (see below), plus a short, focused sanity review if the patch changes behavior or tests. * **Transparency:** the **canonical base prompt lives in-repo** so everyone can inspect and improve it: `ai/rsyslog_commit_assistant/base_prompt.txt` Under the hood, it applies a stable workflow: classify the patch, check for behavior changes, surface concise risks when relevant, and then compose a policy-compliant message. If key metadata (like the non-technical reason or issue URL) is missing, it **asks once** and pauses, exactly as defined in the prompt. ## What it is not * Not required. It is **100% optional**. * Not a gatekeeper or auto-committer. * Not connected to your repository. It only sees what you paste. * **Never paste secrets or tokens.** ## The commit style it reinforces (summary) These rules mirror our **Contribution Guide** and are the source of truth there. * **ASCII only.** * **Title <= 62 chars; body lines <= 72 chars.** * Title format: `<component>: <concise action>` * Start with a brief **non-technical “why”** (e.g., modernization, maintainability, performance, security, Docker/CI readiness, user value). * Add **Impact:** one line when tests or user-visible behavior change. * Include a one-line **Before/After** summary. * Provide a **technical overview** (4–12 lines, conceptual; not line-by-line). * Footer with **full-URL** references: * `Fixes: https://github.com/rsyslog/rsyslog/issues/<id>` when conclusively fixed * otherwise `Refs: https://github.com/rsyslog/rsyslog/issues/<id>` * If an agent helped, add: `AI-Agent: <name version>` See the Contribution Guide for full details and context. ## Example output (style illustration) omhttp: migrate to OMODTX and align retry semantics Why: modernize module internals, reduce edge-case divergence, and improve behavior clarity for high-volume Docker/CI setups. Impact: delivery semantics clarified; docs/tests updated. Before: mixed retry handling; 3xx/5xx behavior not consistently aligned with core defaults. After: 1xx/2xx=success; 3xx=non-retriable; 4xx=non-retriable (overridable); 5xx/transport=retriable by default; uses commitTransaction(). Technical: - switch to OMODTX commitTransaction() interface - align status-code handling with core retry/backoff defaults - keep retryRuleset as optional advanced path - update counters and health behavior; add test coverage Refs: https://github.com/rsyslog/rsyslog/issues/5957 AI-Agent: Codex 2025-06 ## How to use it (2-minute flow) 1. Open: **https://www.rsyslog.com/tool_rsyslog-commit-assistant** 2. Paste a short summary and, if helpful, a small diff or link to it. 3. Provide the **non-technical reason** and **issue URL** if available. 4. Review the proposal, tweak as needed, and paste it into `git commit`. Tip: Keep your pasted diffs minimal. Do not include secrets or credentials. ## Privacy and transparency * You control what you paste. We recommend minimal diffs and no sensitive data. * The exact behavior is documented in the public **base prompt** : **https://github.com/rsyslog/rsyslog/blob/main/ai/rsyslog_commit_assistant/base_prompt.txt** * The Contribution Guide documents the **commit rules** and links to the assistant. ## Feedback and iteration This assistant is **experimental** and will evolve with your input. Please use **GitHub Discussions** to propose rule tweaks, extra checks, or localized guidance: * **https://github.com/rsyslog/rsyslog/discussions** ## Links * Use the assistant: **https://www.rsyslog.com/tool_rsyslog-commit-assistant** * Base prompt (source of truth): **https://github.com/rsyslog/rsyslog/blob/main/ai/rsyslog_commit_assistant/base_prompt.txt** * Contribution Guide and rules: repo `CONTRIBUTING.md` * Feedback: **https://github.com/rsyslog/rsyslog/discussions**

Dogfooding AI in OSS: built the rsyslog Commit AI Assistant — already saves time & avoids wasteful CI runs. 👉 www.rsyslog.com/introducing-the-rsyslog-...

#ai #oss #rsyslog

Introducing the rsyslog commit AI assistant # We are adding a new helper to our Responsible AI First toolbox: the **rsyslog commit (message) assistant**. It is an **optional** ChatGPT Custom GPT that helps contributors write **clear, policy-compliant commit messages** faster. You stay in full control. Nothing in your workflow changes unless you want it to. It follows our responsible “AI First” strategy and is optional, transparent, and keeps the human in full control. * Use it in your browser: **https://www.rsyslog.com/tool_rsyslog-commit-assistant** * Review its exact behavior (full transparency): **https://github.com/rsyslog/rsyslog/blob/main/ai/rsyslog_commit_assistant/base_prompt.txt** * Discuss and give feedback: **https://github.com/rsyslog/rsyslog/discussions** ## Why we built it * **Clarity and consistency:** higher-signal commit history and smoother reviews. * **Lower barrier for newcomers:** fewer style nits, more focus on substance. * **Good AI use:** opt-in, scoped, transparent, and aligned with our contribution rules. * **OSS-friendly cost model:** a Custom GPT lets us support contributors on a lean budget. ## What it does The assistant asks for the essentials and proposes a commit message you can edit: * **Inputs it expects:** a diff (pasted or link), optional issue URL, an optional non-technical reason, and optional context like component or whether a bug was seen in practice. * **What it produces:** a commit message that follows our rules (see below), plus a short, focused sanity review if the patch changes behavior or tests. * **Transparency:** the **canonical base prompt lives in-repo** so everyone can inspect and improve it: `ai/rsyslog_commit_assistant/base_prompt.txt` Under the hood, it applies a stable workflow: classify the patch, check for behavior changes, surface concise risks when relevant, and then compose a policy-compliant message. If key metadata (like the non-technical reason or issue URL) is missing, it **asks once** and pauses, exactly as defined in the prompt. ## What it is not * Not required. It is **100% optional**. * Not a gatekeeper or auto-committer. * Not connected to your repository. It only sees what you paste. * **Never paste secrets or tokens.** ## The commit style it reinforces (summary) These rules mirror our **Contribution Guide** and are the source of truth there. * **ASCII only.** * **Title <= 62 chars; body lines <= 72 chars.** * Title format: `<component>: <concise action>` * Start with a brief **non-technical “why”** (e.g., modernization, maintainability, performance, security, Docker/CI readiness, user value). * Add **Impact:** one line when tests or user-visible behavior change. * Include a one-line **Before/After** summary. * Provide a **technical overview** (4–12 lines, conceptual; not line-by-line). * Footer with **full-URL** references: * `Fixes: https://github.com/rsyslog/rsyslog/issues/<id>` when conclusively fixed * otherwise `Refs: https://github.com/rsyslog/rsyslog/issues/<id>` * If an agent helped, add: `AI-Agent: <name version>` See the Contribution Guide for full details and context. ## Example output (style illustration) omhttp: migrate to OMODTX and align retry semantics Why: modernize module internals, reduce edge-case divergence, and improve behavior clarity for high-volume Docker/CI setups. Impact: delivery semantics clarified; docs/tests updated. Before: mixed retry handling; 3xx/5xx behavior not consistently aligned with core defaults. After: 1xx/2xx=success; 3xx=non-retriable; 4xx=non-retriable (overridable); 5xx/transport=retriable by default; uses commitTransaction(). Technical: - switch to OMODTX commitTransaction() interface - align status-code handling with core retry/backoff defaults - keep retryRuleset as optional advanced path - update counters and health behavior; add test coverage Refs: https://github.com/rsyslog/rsyslog/issues/5957 AI-Agent: Codex 2025-06 ## How to use it (2-minute flow) 1. Open: **https://www.rsyslog.com/tool_rsyslog-commit-assistant** 2. Paste a short summary and, if helpful, a small diff or link to it. 3. Provide the **non-technical reason** and **issue URL** if available. 4. Review the proposal, tweak as needed, and paste it into `git commit`. Tip: Keep your pasted diffs minimal. Do not include secrets or credentials. ## Privacy and transparency * You control what you paste. We recommend minimal diffs and no sensitive data. * The exact behavior is documented in the public **base prompt** : **https://github.com/rsyslog/rsyslog/blob/main/ai/rsyslog_commit_assistant/base_prompt.txt** * The Contribution Guide documents the **commit rules** and links to the assistant. ## Feedback and iteration This assistant is **experimental** and will evolve with your input. Please use **GitHub Discussions** to propose rule tweaks, extra checks, or localized guidance: * **https://github.com/rsyslog/rsyslog/discussions** ## Links * Use the assistant: **https://www.rsyslog.com/tool_rsyslog-commit-assistant** * Base prompt (source of truth): **https://github.com/rsyslog/rsyslog/blob/main/ai/rsyslog_commit_assistant/base_prompt.txt** * Contribution Guide and rules: repo `CONTRIBUTING.md` * Feedback: **https://github.com/rsyslog/rsyslog/discussions**

A new addition to the rsyslog AI agent family. Its' small but pretty useful!

www.rsyslog.com/introducing-the-rsyslog-...

Backticks in RainerScript just got smarter: ${VAR} and adjacent text now work - rsyslog TL;DRBackticks with echo in RainerScript now support brace-style environment variables (${VAR}) and adjacent text (e.g., `echo sasl.password=${KAFKA_PASSWORD}`). This removes a common pitfall when ass...

Improvement on #rsyslog internal environment variable processing: www.rsyslog.com/backticks-in...

Backticks in RainerScript just got smarter: ${VAR} and adjacent text now work **TL;DR** Backticks with `echo` in RainerScript now support **brace-style environment variables** (`${VAR}`) and **adjacent text** (e.g., ``echo sasl.password=${KAFKA_PASSWORD}``). This removes a common pitfall when assembling `key=value` pairs for modules like **`omkafka`**. It’s still a limited, intentional subset—not a full shell. The change was motivated by real-world confusion reported in **issue #5827**. (GitHub) Symbol picture: rsyslog config snippet being shown on a screen. (Image: Rainer Gerhards via AI) ## What changed (and why) Previously, backticks did not recognize `${VAR}`, and unbraced `$VAR` sometimes swallowed trailing punctuation, breaking practical use cases (like `key=value` strings). That led to parsing errors such as “missing equal sign in parameter” when users tried to inject SASL credentials into `omkafka` via env vars. (GitHub) The new logic: * **Understands`${VAR}`** and stops at the closing `}`. * For unbraced `$VAR`, **stops at the first non-`[A-Za-z0-9_]`** character and emits that character literally, so patterns like `$VAR!` work as expected. * Remains intentionally small: no general shell evaluation (e.g., **no** `$(...)`). That’s by design and documented. (rsyslog.com) Result: patterns admins expect from daily shell usage now behave sensibly inside RainerScript backticks—without turning backticks into a shell. (rsyslog.com) ## Practical example: `omkafka` SASL password from an env var **Before (clunky workaround)** You often had to pre-compose the full `key=value` into an environment variable and then expand it: # in a secure environment file or service unit export SASL_PWDPARAM='sasl.password=${KAFKA_PASSWORD}' export KAFKA_PASSWORD='supersecret' # rsyslog.conf action( type="omkafka" broker=["kafka.example.com:9093"] confParam=[ "security.protocol=SASL_SSL", "sasl.mechanism=SCRAM-SHA-512", "sasl.username=myuser", `echo $SASL_PWDPARAM` ] ) Users hit errors like “missing equal sign” while trying variations around this approach. (GitHub) **Now (direct and clear)** Compose the `key=value` inline with `${…}`: export KAFKA_PASSWORD='supersecret' action( type="omkafka" broker=["kafka.example.com:9093"] confParam=[ "security.protocol=SASL_SSL", "sasl.mechanism=SCRAM-SHA-512", "sasl.username=myuser", `echo sasl.password=${KAFKA_PASSWORD}` ] ) This matches everyday expectations while staying within RainerScript’s documented backtick subset. For reference, see the `omkafka` and RainerScript docs. (rsyslog.com) ## Availability * **Merged:** **August 15, 2025**. * **Daily stable build:** **August 16, 2025**. * **Scheduled stable build:** **8.2508.0** later in August. If you use the daily stable channel, you can pick it up right away; otherwise you’ll get it with **8.2508.0**. ## Security & ops notes * Prefer **environment files** (e.g., systemd `EnvironmentFile=`) with strict permissions over hard-coding secrets in `rsyslog.conf`. * Remember that process environments are visible to privileged users via `/proc`; secure the host accordingly. * This change is **backward-compatible** , fixing previously broken or surprising edge cases. ## Background & related reading * **Issue #5827 –** _omkafka: unable to use environment variable to password`sasl.password`_ (parse error context and motivation). (GitHub) * **RainerScript string constants** – canonical reference for backticks and their intentional limitations. (rsyslog.com) * **`omkafka` module docs** – configuration parameters and examples. (rsyslog.com) If you previously worked around this with pre-composed env vars, you can simplify your configs now. And if you see anything unexpected, please share a minimal snippet—reports like **#5827** help us make rsyslog better for everyone. (GitHub)

Improvement on #rsyslog internal environment variable processing: www.rsyslog.com/backticks-in-rainerscrip...

Shipping Better Docs with AI: Restructuring Module Parameters for Clarity and Consistency - rsyslog As you may know from past articles, we’re in the process of a major documentation overhaul. I wanted to share a focused task we’ve been working on over the past couple of days. This effort highlights ...

We're using a responsible AI-First approach to overhaul rsyslog's documentation. Our 3-stage, human-in-loop workflow ensures accuracy and consistency. We use AI to generate and validate changes, all guided by human expertise.

#rsyslog #opensource #promptengineering

Original post on mastodon.social

We're using a responsible AI-First approach to overhaul rsyslog's documentation. Our three-stage, human-in-the-loop workflow ensures accuracy and consistency. We use AI to generate and validate changes, all guided by human expertise. This creates better docs for both users and AI tools.

Read […]

I have just updated the #rsyslog Assistant (#AI) to ChatGPT 5 model. We assume this will further improve it's reliabiability, but I will also carry out tests against known real-world question scenarios.

Original post on mastodon.social

Remember when Wikipedia was dismissed as unreliable and Britannica was the gold standard? Today, Wikipedia is the default, and AI often serves its content directly.

The same transformation is coming for software docs, development, and support. With rsyslog, we are building this AI-first future […]

AI is not for the lazy ones - AI prompt for doc work as an example - Rainer Gerhards It was interesting to see the commotion and rejection our ‘AI First’ announcement for rsyslog caused. I thought that was a very high level posting detailing that we worked for … Continue reading "AI i...

#AI is only as good as humans are ready to invest effort. Based on the responses I got on #rsyslog #AIFirst, this seems to be unkonwn of many. So I thought I show how far we to in trying to do things right - meet our base prompt for rsyslog doc rewrite project.

rainer.gerhards.net/2025/07/ai-i...

Original post on mastodon.social

The panic around our #AIFirst vision was quite entertaining as it showed people are just to lazy to fully make a judgement call. Let me show as an example our current base prompt for the doc improvement custom GPT and some more background on our thinking […]

Original post on mastodon.social

Obviously, there is some debate on #rsyslog "AI First" initiative.

#AI done right is a great tool. Don't think we are foolish enough to go lazy. Some more details are in this article.

Don't be lazy yourself and read it before doing your judgement ;-) […]

Evolving rsyslog Documentation with AI: From Chaos to a Plan - Rainer Gerhards TL;DR: The rsyslog documentation was a mess — and I knew it. It’s now only “partly OK”, but for the first time we have a clear structure, concept, and plan. … Continue reading "Evolving rsyslog Docume...

If it is slop or not depend on human intent. If the intent is to be lazy, you'll most probably get garbage. If the intent is to do quality work and invest effort, good results are likely.

As an example, see my work on the #rsyslog doc. This will be a long effort and require human/AI collaboration.