Recording rules is a clever concept introduced by Prometheus for storing results of query expressions in a form of a new time series. It is similar to materialized view and helps to speed up queries by using data pre-computed in advance instead of doing all the hard work on query time. Like materialized views, recording rules are extremely useful when user knows exactly what needs to be pre-computed. For example, a complex panel on Grafana dashboard or SLO objective.

Share: When we posted our first ever Momentum blog about a year ago detailing our 2021 achievements, we were just weeks away from Russia’s renewed attack of Ukraine. While the war isn’t won yet and we’re approaching the one year anniversary of the attack, it’s heartening to see how much has changed around the world and that almost everyone now knows the expression: Slava Ukraini! So if we had to choose one word to best describe 2022 it might be: Resilience.

Share: One of the latest benchmarks we did was for OSMC 2022 talk VictoriaMetrics: scaling to 100 million metrics per second - see the video and slides. While the fact that VictoriaMetrics can handle data ingestion rate at 100 million samples per second for one billion of active time series is newsworthy on its own, the benchmark tool used to generate that kind of load is usually overlooked. This blog post explains the challenges of scaling the prometheus-benchmark tool for generating such a load.

VictoriaMetrics is proud to announce that we consider vmbackup and vmbackupmanager to be feature-complete solutions as of release 1.85.3. These backup components are essential for ensuring the safety and integrity of your data, and we have made a number of improvements in recent releases to make them even more reliable and user-friendly.

In monitoring, the term cardinality defines the number of unique time series stored in a time series database (TSDB). The higher the cardinality, the more resources are usually required for metrics processing and querying.

Share: VictoriaMetrics is a monitoring solution. It was designed to collect and process telemetry from many systems, provide a retrospective view, and forecast metrics for capacity planning. But what about monitoring VictoriaMetrics itself? There is one of the software development approaches called Observability Driven Development (ODD). In a nutshell, it means that developers should always keep in mind that software needs to be transparent to the person who uses it. Does your software make backups?

Grafana Labs Mimir is a new time series database under AGPLv3-license. The engineering team did a great job by taking the best from Cortex TSDB, reducing its complexity and improving scalability in the same time. According to tests by Grafana Labs, Mimir can scale to a billion active time series and 50 million samples/s ingestion rate.

We need to remove a vmstorage node from VictoriaMetrics cluster gracefully. Every vmstorage node contains its own portion of data and removing the vmstorage node from the cluster creates gaps in the graph (because replication is out of scope).

When looking for a highly scalable time series database, there are a number of criteria to investigate and evaluate. First up, it’s always a good idea to consider open source software. It’s more likely to have gone through comprehensive troubleshooting, it’s typically more reliable as it has more timely and widespread peer-review, it better guarantees technology independence, it’s easier to find engineers who are familiar with it and it has great security.

vmagent supports both the pushing and pulling (scrape) of metrics and here are examples of high availability setups for both cases.