So are we doing Tempo, Victoria, Influx... what for home lab observability now?
#observability
0 posts0 participants0 posts today
Yowza. I got a nice boost from @nixCraft and my instance survived. That was a helluva boost.
I recently made two big changes to my observability stack:
- I realized that I don't need to run 3 pods of Prometheus, each taking up about 4GB of ram. This cut resource consumption to 1/3.
- Switched to Deckhouse Prom++, which cut the single instance to about 1/10.
New Open-Source Tool Spotlight 
Kubernetes History Inspector (KHI) is an agentless log viewer built for visualizing Kubernetes audit logs. Its timeline-based log analysis and resource relationship diagrams simplify cluster troubleshooting—no complex setups or commands required. #Kubernetes #Observability
Project link on #GitHub 
https://github.com/GoogleCloudPlatform/khi
#Infosec #Cybersecurity #Software #Technology #News #CTF #Cybersecuritycareer #hacking #redteam #blueteam #purpleteam #tips #opensource #cloudsecurity
— 
P.S. Found this helpful? Tap Follow for more cybersecurity tips and insights! I share weekly content for professionals and people who want to get into cyber. Happy hacking 
ClickStack – open-source Datadog alternative by ClickHouse and HyperDX
https://clickhouse.com/use-cases/observability
#HackerNews #ClickStack #ClickHouse #HyperDX #open-source #Datadog #alternative #observability
ClickHouseClickStack: High-Performance Open-Source Observability | Logs, Metrics, Traces with ClickHouseClickStack is a high-performance, open-source observability stack powered by ClickHouse. Unify logs, metrics, traces, and session replays with lightning-fast queries and 10-100x cost savings on even your highest-cardinality data.
Hear me out - one REALLY BIG grafana for the EU
Is Node.js the future of backend development, or just a beautifully wrapped grenade?
Lately, I see more and more backend systems, yes, even monoliths, built entirely in Node.js, sometimes with server-side rendering layered on top. These are not toy projects. These are services touching sensitive PII data, sometimes in regulated industries.
When I first used Node.js years ago, I remember:
• Security concepts were… let’s say aspirational.
• Licensing hell due to questionable npm dependencies.
• Tests were flaky, with mocking turning into dark rituals.
• Behavior of libraries changed weekly like socks, but more dangerous.
• Internet required to run a “local” build. How comforting.
Even with TypeScript, it all melts back into JavaScript at runtime, a language so flexible it can hang itself.
Sure, SSR and monoliths can simplify architecture. But they also widen the attack surface, especially when:
• The backend is non-compiled.
• Every endpoint is a potential open door.
• The system needs Node + a fleet of dependencies + a container + prayer just to run.
Compare that to a compiled, stateless binary that:
• Runs in a scratch container.
• Requires zero runtime dependencies.
• Has encryption at rest, in transit, and ideally per-user.
• Can be observed, scaled, audited, stateless and destroyed with precision.
I’ve shipped frontends that are static, CDN-delivered, secure by design, and light enough to fit on a floppy disk. By running them with Node, I’m loading gigabytes of unknown tooling to render “Hello, user”.
So I wonder:
Is this the future? Or am I just… old?
Are we replacing mature, scalable architectures with serverless spaghetti and 12-factor mayhem because “it works on Vercel”?
Tell me how you build secure, observable, compliant systems in Node.js.
Genuinely curious.
Mildly terrified and maybe old.
Save the date! After #GrafanaCON Day 1, join Embrace, Tailscale & Grafana for happy hour at SPIN Seattle!
Drinks, bites & ping pong — May 6, 4–6 PM.
Space is limited — RSVP now!
Excited to kick off Season 3 of my #ITOps Query #podcast with guest @mipsytipsy. Topics included #AIobservability, #observability -driven development, "Observability 2.0," and Majors' warning / call to action for SREs about the coming wave of #AIagents: "I don't want to see happen to #SRE what happened to QA 15 years ago." #o11y #agenticAI #softwaredevelopment
Man Prometheus is a pain to recover once its data store is in any way out of shape. Did NOT help that it was buried inside Kubernetes inside a PVC.
Thankfully it was only Dev environment today but if this ever pages on Prod we're losing data as it stands.
I'll write something up for a run book but eesh.
Announcing our new EU data region! 
If you want to use Honeybadger but must keep your data outside the U.S. (you know, for whatever reason) — this is for you.
Read the blog post, and/or stick around for the key takeaways. 
https://www.honeybadger.io/blog/eu-data-residency/?utm_source=mastodon&utm_medium=social
Honeybadger Developer BlogHoneybadger’s EU data center is now generally available 🇪🇺You can now host your Honeybadger data outside the U.S. to meet EU data residency requirements.
Replied in thread
@gwensnyder.bsky.social it might be on #DarkSocial to elide #observability
Replied to paigerduty
@paigerduty Love it!!! And really nice because the postal service serves as a metaphor in many other cases, which benefits knowledge transfer!
So one that I like to use is to illustrate how we are more interested in useful dashboards than creating a dashboard(s) for everything.
We sometimes market this as "the single pane of glass solution". So I encourage people to think about a stained glass window instead.
Single panes of glass always look the same. Homogenous and uncolored from any angle or perspective, under most any light. Single panes of glass can be broken, they have no internal structure.
Stained Glass looks different if you change your perspective only slightly. Light is refracted in multiplicative ways through the different colors of glass. Different kinds of light have different shades and hues. Most importantly, stained glass pieces are deliberate, they depict something. These windows can only be broken in sections, which is a nice example of systems thinking as well!
personally I love a metaphor when introducing and explaining tech stuff, esp in the world of #o11y where terms are often overloaded, misunderstood, misapplied
so from a thread I spawned last year trying to suss out wtf a telemetry pipeline was I landed on this postal service metaphor 
(elaborated on here: https://thenewstack.io/telemetry-pipelines-collectors-and-agents-whats-the-difference/)
lmk what ya think or send me your fave #observability metaphors
Replied in thread
@fosdem Opening the #FOSDEM #Monitoring & #Observability devroom on Sunday. It was empty in the morning but flipped to being full with lines outside soon enough.
Again, another videobox, showing the built-in display
Edit: the first photo was from 30 minutes before we opened. During opening, we were at maybe 1/3th capacity already
OpenTelmetry tracing views of CI including inside jobs is pretty cool
(it's at the bottom)
https://www.honeycomb.io/blog/most-important-developer-productivity-metric-build-times
Honeycomb · The Most Important Developer Productivity MetricAt Honeycomb, we believe in ensuring an awesome developer experience for our engineers, so we've tried to keep our build times under 15 mins.
I think I'm going to focus next year on #observability in context of better debugging and improving finding problems and testing more in the sense of probing.
So Otel and observability are having a moment, is there anything that does that for network flows and packets? Like ships traces from interfaces into signoz / LGTM etc?