Prometheus 3.0.0 Beta was announced during PromCon EU 2024 last week. While it’s not ready for production users, you can already see the features it brings. They include:

- A brand new UI enabled by default. It has a modern look & feel based on Mantine UI with light and dark modes, a face-lifted menu structure, a metrics and labels explorer, a PromLens-style query tree view, and a query explanation tab.
- Remote Write 2.0 (we covered it before in this digest).
- OpenTelemetry support with UTF-8 characters for metric and label names and the ability to serve as a native receiver for the OTLP Metrics protocol.
- Native histograms, which are a higher efficiency and lower cost alternative to classic histograms.
- Agent mode is declared stable.

Find more information via the following resources:
- Release announcement in the project's blog
- GitHub release
- “A Look at the New Prometheus 3.0 UI” by Julius Volz, PromLabs

#news #releases #observability

Wireshark Foundation has introduced Stratoshark created by Sysdig and advertised as "Wireshark for the Cloud".

Stratoshark is a tool that provides deep visibility into application-level behaviour by analysing cloud system calls and logs. It is built on the legacy of Wireshark and Falco, designed for Cloud Native environments, and supports the same file format as Falco and Sysdig CLI.

- Website
- LinkedIn announcement
- “Troubleshooting CrashLoopBackOff with Stratoshark”

#news #tools #security #observability

Kubernetes History Inspector (KHI) from Google Cloud is Open Source now.

As its description states, “KHI transforms vast quantities of logs into an interactive, comprehensive timeline view.” This tool collects the Kubernetes logs and visualises them providing SREs with an interactive Web-based GUI.

KHI displays the status of resources on a timeline, correlates various types of logs (event logs, audit logs, network, etc.) to help you find their relationships and dependencies, and allows you to interactively apply various filters. Currently, the Kubernetes clusters in Google Cloud (i.e. GKE) are supported only, with more options to follow.

▶️ GitHub repo

#tools #GCP #observability #news

Do you validate your Prometheus and other observability-related rules? Sometimes, you want to check them thoroughly and reviewing their syntax is not enough. Here’s when this tool comes in handy.

Promruval (Prometheus Rule Validator) validates rules metadata and expression properties against your actual setup and requirements. To benefit from it, you’ll need to create a YAML configuration defining allowed values, limits, and other constraints and then invoke promruval validate in your CI pipeline. Some of the tool’s highlights are:

- Support for Prometheus, Thanos, Mimir, and Loki rules formats.
- Support for rule files in YAML and Jsonnet formats.
- A huge list of available validators, which cover rule groups, labels, PromQL & LogQL expressions, and more.
- An ability to have multiple configuration files.

Language: Go | License: Apache 2.0 | 140 ⭐️

▶️ GitHub repo

#tools #observability

Promtail is now deprecated

Last week, Grafana Loki 3.4 was announced. This release's highlights included “Promtail merged into Alloy”:
- Promtail is a lightweight agent shipping logs to Grafana Loki.
- Grafana Alloy is a distribution of the OpenTelemetry Collector with built-in Prometheus pipelines and support for metrics, logs, traces, and profiles.
- Merging Promtail into Alloy means the former is officially declared deprecated. The project will reach its EOL on March 2, 2026. A guide on migrating from Promtail to Grafana Alloy is available.

#news #releases #observability

A couple of recent articles on optimising memory consumption in Prometheus:

1. “Prometheus: How We Slashed Memory Usage” (Devoriales).

“In the production Kubernetes cluster I worked on, Prometheus memory usage climbed past 55 GB, peaking at 60 GB, despite an already oversized node. Indeed, the environment was rapidly growing in the number of applications, but the situation was still not sustainable.”

2. “Understanding and optimizing resource consumption in Prometheus” (Palark blog).

“While Prometheus is an excellent and capable monitoring system, one aspect I find very frustrating is its resource consumption. If this frustrates you as much as it does me, let’s break down the causes of this issue and see how to address it.”

#articles #observability

This project is a Kubernetes operator that implements a so-called health-check platform.

Canary checker is a monitoring tool that includes 35+ ready-to-use health checks for various Kubernetes workloads. It works as a K8s operator that has no dependencies and offers:

- checks for popular protocols (HTTP, DNS, ICMP, TCP), databases (SQL-based, MongoDB, Redis, Elasticsearch), infrastructure components (such as Kubernetes resources and EC2 instances), file systems (NFS, S3, SMB), and more;
- alert aggregation from Prometheus, AWS CloudWatch, and Dynatrace;
- integration testing support for JMeter, JUnit, K6, Newman and Playwright;
- Prometheus exporter replacement with custom metrics from the check’s results;
- built-in Web dashboard displaying existing checks and their stats.

▶️ GitHub repo

Language: Go | License: Apache 2.0 | 326 ⭐️

#tools #observability

OpenTelemetry got a one-command Linux setup

The OpenTelemetry Packaging SIG introduced a ready-to-use opentelemetry package for Linux distributions that installs the OpenTelemetry Injector together with the OpenTelemetry SDKs and auto-instrumentation packages for various programming languages (Java, .NET, Node.js, and Python).

It is now available as an APT repository (for Debian, Ubuntu, and derivatives) and a YUM repository (for Fedora, RHEL, and derivatives). Find more details in this announcement or read the full instructions on GitHub.

#news #observability