Coding @VictoriaMetrics. Learning Observability. I contribute bugs to VictoriaMetrics.
VictoriaMetrics Playground
Joined July 2015
- Tweets591
- Following57
- Followers523
- Likes810
Jiekun retweeted
Hallo, München! 🇩🇪
We're only 1 week away from @PromConIO 2026, the biggest conference for #Prometheus fans!
📍 Google Offices, Munich, Germany
🗓️ October 7–8
We’re looking forward to connecting with the community, showcasing our latest innovations, and discussing everything from #metrics, #logs, and #traces to the latest in #OpenSource #observability.
And don’t miss this talk:
🎤 Tracking Cardinality Explosions In Flight
Presented by our Co-founder, @hagen1778, and Langson Zhang, Software Engineer at Roblox.
🗓️ October 8 at 12:15 PM CET
hubs.li/Q04xbwlc0
This talk is about estimating cardinality in flight, while the metrics are still streaming in. The estimator runs as a separate pipeline alongside the storage, not inside it. A small estimator maintains a HyperLogLog sketch for each label dimension and publishes the results as standard metrics in the Prometheus exposition format. It receives data over remote-write and is scraped like any other target. There are no queries against storage, no new query language to learn, and nothing tied to a specific vendor, so any Prometheus-compatible setup can use it.
📌 Save your seat and come say hi in Munich. See you at #PromCon! 🎉
hubs.li/Q04xbwj80
#VictoriaMetrics #MonitoringSolutions #TechEvents #Prometheus
Jiekun retweeted
Hallo, München! 🇩🇪
#VictoriaMetrics is thrilled to be a Diamond Sponsor at @PromConIO, the biggest conference for #Prometheus fans!
📍 Google Offices, Munich, Germany
🗓️ October 7–8
We look forward to connecting with the community, showcasing our latest innovations, and discussing everything from #metrics, #logs, and #traces to the latest in #OpenSource #observability.
hubs.li/Q04xlHSb0
And don’t miss this talk:
🎤 Tracking Cardinality Explosions In Flight
Presented by our Co-founder, @hagen1778, and Langson Zhang, Software Engineer at Roblox.
🗓️ October 8 at 12:15 PM CET
hubs.li/Q04xlHVq0
This talk is about estimating cardinality in flight, while the metrics are still streaming in. The estimator runs as a separate pipeline alongside the storage, not inside it. A small estimator maintains a HyperLogLog sketch for each label dimension and publishes the results as standard metrics in the Prometheus exposition format. It receives data over remote-write and is scraped like any other target. There are no queries against storage, no new query language to learn, and nothing tied to a specific vendor, so any Prometheus-compatible setup can use it.
📌 Save your seat and come say hi in Munich. See you at #PromCon! 🎉
#VictoriaMetrics #MonitoringSolutions #TechEvents
Jiekun retweeted
I like seeing real-world comparisons of various tools we use. Most log stack comparisons read like vendor pitches, so it's nice to see one grounded in actual production workloads. This one runs Loki and VictoriaLogs side by side on 500 GB of logs over 7 days, on identical hardware.
There are a handful of tools in this space, and Loki and VictoriaLogs are two of the more popular ones. The test results here are very interesting and I would like to try a similar comparison in my own environment. Harshit Luthra has put together a balanced comparison and outlines when each option is the right call, especially for label-heavy workloads and Grafana-native shops. Check it out!
lckhd.eu/MchTXw
#VictoriaLogs #Loki #Grafana
Jiekun retweeted
📢 April episode is out! 📢
#VictoriaMetrics: From Monitoring to Observability
With @horovits hosting @hagen1778, Co-Founder and Engineering Manager @VictoriaMetrics
Listen on your favorite platform:
🎧 creators.spotify.com/pod/pro…
📺 youtube.com/watch?v=iBa-Zcr8…
Jiekun retweeted
Airbnb migrates its high-volume metrics pipeline to OpenTelemetry, moving from StatsD and Veneur to OTLP, the OpenTelemetry Collector, and VictoriaMetrics vmagent. The new stack ingests 100M+ samples per second in production. Read the full story here: ift.tt/M1To0fs
Jiekun retweeted
One of the most thorough #Logs shippers benchmarks I've seen. Covering Vector, Promtail, Grafana Alloy, Fluent Bit, OpenTelemetry Collector, and more.
A benchmark suite, steps to reproduce, Grafana snapshots - all is there.
Well worth a read: victoriametrics.com/blog/log…
Jiekun retweeted
kinda crazy how VictoriaMetrics/VictoriaLogs came out of nowhere and got popular. I have no idea where they rank as an APM. what are primary reasons people are choosing this solution? is there a better scalability story?
Jiekun retweeted
Very clear explanation on how Go memory allocator works.
New article: Go's Memory Allocator 🚀
How Go manages heap memory for thousands of goroutines. 64MB arenas → 8KB pages → spans → fixed-size slots. A three-level cache where the fast path is completely lock-free.
👉 internals-for-interns.com/po…
#golang #go #runtime #programming
Jiekun retweeted
The video recording of this talk - youtube.com/watch?v=zO1qvQIb…
Jiekun retweeted
If you are interested in the history of VictoriaMetrics company, then read the following article - underdogfounders.substack.co…
It contains a few small inaccuracies, but overall the article is great!
Jiekun retweeted
VictoriaMetrics v1.132.0 contains an optimized OpenTelemetry protocol parser, which is up to 6x faster than the previous one.
I don't like OpenTelemetry protocol for metrics because it is over-engineered and bloated, but it was fun optimizing the OTEL metrics' parser with the help of github.com/VictoriaMetrics/e… package. I suspect the resulting parser is the fastest one for OTEL protocol for metrics.
If you are interested in details, then take a look at the commit github.com/VictoriaMetrics/V…
Jiekun retweeted
This benchmark shows that VictoriaLogs delivered ~94 % lower query latency, ~40 % less storage usage, and used less than half the CPU and RAM compared to Loki on a 500 GB/7-day log workload
➜ ku.bz/PVljlj8Ks
Jiekun retweeted
Because i use my brain and know the story. This is not a vibed story by Gaurav. But it also misleading.
They had CPU bottleneck due wrong architecture in GO.
The rewrite could have be a fix in the go code.
Rust just makes it difficult to write the wrong arch.
Jiekun retweeted
$2800/month with GC is about 500k series ($6.5 per 1k series according to their pricing). This load seems very small and I don't get it why it will be $800 for self-hosted. Btw, in @VictoriaMetrics cloud it will cost $200/month.
Jiekun retweeted
One of the top-selling features of OpenTelemetry is easy vendor switch. This is a false promise for metrics. Every vendor applies its' own set of transformations over the metric names and labels ingested over OTEL format before storing the ingested metrics, because of over-complicated OTEL format for metrics. This makes it very hard to migrate between vendors, since you'll need to rewrite all the queries over OTEL metrics during the migration.
The same issue exists for logs. For example, some systems (such as Loki and VictoriaLogs) support the log stream concept (e.g. logs received from a single application instance). They can naturally map resource attributes from OTEL format for logs to log stream labels. Other systems (such as ElasticSearch) do not support log streams, so they map resource attributes to some other log fields with some prefixes. This makes it almost impossible to migrate between vendors for logs.
Which other features in OTEL convince you to migrate to it from some other well-supported formats for metrics and logs?
Jiekun retweeted
Day 1 of #PromCon was amazing! 🚀
Lots of great talks, great people, and even better #observability stories.
See you tomorrow. Drop by our table for swag and conversations about open source observability solutions! #prometheus #opensource #monitoring
Jiekun retweeted
Huge thanks to @victoriametrics for sponsoring #PromCon 2025! 🙏 Your support helps bring the Prometheus community together to share knowledge and push observability forward. 🚀
Be sure to checkout their booth next week!
Jiekun retweeted
The search on the VictoriaMetrics docs just got a big upgrade.
Give it a try: docs.victoriametrics.com
Jiekun retweeted
New blogpost alert: Alerting Best Practices victoriametrics.com/blog/ale…
