PromCon EU 2025
who it's for
Prometheus users and operators, observability engineers, platform engineers, site reliability engineers, monitoring engineers, Grafana dashboard developers, OpenTelemetry engineers
topics
native Prometheus instrumentation; OpenMetrics 2.0 text exposition; Prometheus downsampling techniques; Prometheus agent deployment patterns; OpenTelemetry delta temporality support; native histograms in Prometheus text; PromQL usage and proxies; SAAFE alert prioritization model; Prometheus with Parquet long-term storage; Perses Go SDK for dashboards
| title | author | length |
|---|---|---|
| Welcome | Richard Hartmann | 07m |
Welcome notes focus on logistics not content: Wi‑Fi at the entrance and sign‑in details, repeated microphone problems during setup, and a live stream so… | ||
| Why I Recommend Native Prometheus Instrumentation over OpenTelemetry | Julius Volz | 28m |
Recommendation: use Prometheus client libraries for Prometheus metrics because native instrumentation preserves service discovery + pull semantics and… | ||
| Everything you need to know about OpenMetrics 2.0! | György Krajcsovits, Bartłomiej Płotka | 30m |
OpenMetrics 2.0 proposes a writable, UTF‑8 text exposition that encodes complex types so you can see and query the same thing: it adds native histograms in… | ||
| Downsampling in Prometheus | Jan Fajerski | 24m |
Downsampling is defined as reducing sample rate while keeping recent full-resolution data; current practical options are federation or recording rules with… | ||
| Auto-generated coherent Observability | Arthur Silva Sens, Dominik Süß | 30m |
Weaver takes OpenTelemetry-style YAML schemas and generates type-safe Go clients (no more string labels), Prometheus-compatible metrics, autogenerated docs… | ||
| No Jsonnet? No Problem! Prometheus in Perses, Powered by Go | Saswata Mukherjee, Hélia Barroso | 25m |
Demonstrates a working Perses pipeline that replaces Jsonnet dashboards with code: a Perses Go SDK that builds dashboard structs, a PromQL builder that… | ||
| OTEL delta temporality support in Prometheus | Fiona Liao | 28m |
Prometheus accepts OpenTelemetry delta metrics natively behind a feature flag instead of rejecting or converting them to cumulative, which avoids out‑of‑order… | ||
| Eight definitions of “observability” vs. “monitoring” – and why none of them work | Björn Rabenstein | 36m |
Eight proposed distinctions get evaluated with a simple CUP rating (correctness, usefulness, popularity) and none pass without flaws: the control theory… | ||
| From Manual to Managed: Prometheus Agent Deployment at Scale | Mihail Mihaylov | 28m |
The team deploys Prometheus agents four ways and explains tradeoffs: Helm + ArgoCD for seconds-scale installs but requires Kubernetes and cluster management… | ||
| SAAFE - A prioritized alerting model to troubleshoot your incidents | Jorge Creixell, Manoj Acharya | 22m |
Grafana Labs describes a working, open-source alerting pipeline that scores and ranks alerts by SAAFE categories (Saturation, Amend, Anomaly, Failure, Error)… | ||
| A History of Automatic Aggregations | Raphaël Bizos | 19m |
Criteo runs an automatic aggregation layer that drops instance labels and applies counters rates, gauge percentiles and histogram merges to cut cardinality and… | ||
| Building an Observability Agent for Rapid Root Cause Analysis using Prometheus metrics | P. Yekbote | 32m |
A Prometheus-first observability agent is built that queries Prometheus metrics, pulls logs and traces from OpenSearch via Prometheus MCP server and OpenSearch… | ||
| Engage Prometheus in Rust Ecosystem | Ruihang Xia | 29m |
Rust libraries provide a native PromQL parser (reusing the Prometheus YAK grammar) and implementations of Prometheus remote protocols, used by hundreds of… | ||
| Beyond TSDB: Unlocking Prometheus with Parquet for Modern Sc... | J. Vazquez, A. Protasio, M. Hoffmann | 29m |
Projects Mimir, Cortex and Thanos replace Prometheus TSDB with Parquet columnar storage, storing one series per row and sharding chunks into 8‑hour chunk… | ||
| Alertmanager Has Amnesia – Should We Fix It? | Joel Verezhak | 39m |
Alertmanager loses in-memory alert store on reboot by design, but DNS drops and clustered replicas cause missed inhibitions and duplicate notifications. Using… | ||
| CPU Hardware Counters | Bryan Boreham | 33m |
CPU chips expose a PMU with cache, branch and TLB counters accessible via Linux perf but cloud VMs typically hide them so you need a metal/whole‑chip instance… | ||
| Reverse-Engineering PromQL Usage: A Proxy’s Tale | Nicolas Takashi | 29m |
Built proxy sits between clients and any Prometheus-compatible backend and forwards queries while asynchronously persisting raw PromQL, query type/step, label… | ||
| OpenTelemetry developments in Prometheus over the last year | Arve Knudsen, Owen Williams | 28m |
Prometheus ships an OTLP ingestion endpoint with concrete changes: since 3.4 it offers a no-translation option to keep OpenTelemetry metric and label names… | ||
| UX Research Report: Prometheus and OTel's Resource Attributes. | Victoria Nduka, Andrej Kiripolský | 29m |
UX research via CNCF LFX mentors collects feedback from 70 users surveyed (7 interviews, 63 survey respondents) and identifies four workflows: mapping all OTel… | ||
| From Logs to Metrics: OpenSearch's Prometheus support with Agentic AI and... | A. Jadhav, K. Hounslow | 25m |
OpenSearch has a merged backend plugin merged for native Prometheus ingest while the front-end metrics UX remains in progress; experimental features include an… | ||
| Debugging the Pipeline: Observability for Your Observability with OTel + Prometheus | Israel Blancas | 25m |
Shows how to monitor OpenTelemetry Collector + Prometheus with concrete signals: exporter queue length, Q sizes and send_fail counters, process CPU/memory… | ||
| Counting What You Care About in Your Security Data Pipeline | Attila Szakács | 23m |
AxoSyslog is a high-performance fork of syslog-ng that replaces CSV blobs with native Prometheus metrics and adds programmable, per-log metric emission via… | ||
| Lightning Talk: Schema Inference | Nicolas Takashi, Arthur Silva Sens | 05m |
A small tool named Teleat extracts runtime telemetry (OTLP or scraped Prometheus /metrics), infers metric names types labels and label values, and produces… | ||
| Lightning Talk: PHP and Prometheus For better Metrics | Syed Usman Ahmad | 05m |
Shows a working Prometheus exporter that scrapes live PHP info metrics and sends them to Prometheus, demoing a WordPress on Docker deployment where a community… | ||
| Lightning Talk: Building Prometheus for Platform Engineering | Virginia Diana Todea | 04m |
Quick, humorous lightning riff that tries to define platform engineering by showing a GPT-generated blueprint: it lists concrete components like scraping and… | ||
| Lightning Talk: Prometheus Rules management and validation | Hervé Nicol | 04m |
Prometheus rules live in a git repo, get packaged as a Helm chart, and Flux deploys them to clusters; alerts must keep specific labels and annotations such as… | ||
| Lightning Talk: How to detect and delete unused metrics? | Aliaksandr Valialkin | 05m |
Unused metrics consume memory and disk; two practical detection methods are proposed: the manual method that pulls all metric names from… | ||
| Lightning Talk: ObservabilityWithout Monitoring - A Thought Experiment | Igor de Sousa | 06m |
Argues monitoring doesn’t always yield observability: black‑box cars and electric models hide telemetry so users and mechanics rely on slow, noisy signals and… | ||
| Lightning Talk: Prometheus Bug Scrub | Bryan Boreham | 06m |
Prometheus has 221 open pull requests and 522 open issues; a weekly Tuesday bug scrub appears on the Prometheus.io calendar where a volunteer shares a screen… | ||
| Lightning Talk: Scrape Trolley Dilemma | Bartek Protka | 05m |
Prometheus and OpenTelemetry collectors hit memory limits when a single scrape brings high cardinality series and the common outcome is either a crash loop or… | ||
| Lightning Talk: Anomaly Detection - Robust Strategy | Jorge Creixell | 04m |
Algorithm builds prediction bands from a median over one day and an approximate median absolute deviation (MAD) times 2, with an experimental mad_over_time for… | ||
| Lightning Talk: Instant Vectors ≠ Scalars - How it bit me | Julius Rickerti | 05m |
Graphing a blackbox exporter across locations with Victoria Metrics triggers a surprising behavior: a binary-operator right hand side label push down applies… | ||
| Lightning Talk: How We Validate User Initiated Config Changes in Promet... | G. Michalec, I. Skorjanc | 05m |
A small Go tool called PR checker runs on every pull request in a single large git repo, parses Prometheus rules with the Go Prometheus parser, and enforces… | ||
| Lightning Talk: Practical OpenTelemetry with Prometheus 3.0 | Arve Knudsen | 05m |
Prometheus adds an info promql function that joins metrics with the target_info metric by default, pulling OpenTelemetry resource attributes (for example the… | ||
| Lightning Talk: Managing Silences - Provisioning Silences into Alertmanager | Théo Brigitte | 04m |
Company runs HA Alertmanager per Kubernetes cluster and manual silences across many clusters proves unscalable and error prone (engineers set expiries to year… | ||
| Lightning Talk: Prometheus adoption stats | Richard Hartmann | 02m |
Annual Grafana-sourced adoption numbers show Prometheus adoption 1.4 (single public exception to otherwise guarded data), with year-over-year growth that… | ||
| Lightning Talk: Would you like your rate smoothed or anchored? | Björn Rabenstein | 05m |
Prometheus adds two new experimental rate functions: anchored and smooth. Anchored computes integer deltas using the last sample before and at the range so it… | ||
| Lightning Talk: Improved Prometheus filtering in Grafana | Victor Marin | 05m |
Compact UI replaces large filter bars with pill-style labels and full value lists on hover, plus improved keyboard navigation. Dashboard default filters are… | ||
| Lightning Talk: Correlating Time Series at Scale | Kathrin Paschen | 06m |
Implements a published algorithm that previously scaled to 5,000 series and runs it against a Kubernetes testbed that produces 40–80,000 time series… | ||
| Lightning Talk: Alert Quorum Universal Aggregator - AQUA | Mirek Chocholous | 07m |
AQUA polls and fetches multiple Alertmanager instances across clusters, computes a quorum-based suppression so alerts only notify when multiple regions agree… | ||
| Lightning Talk: Prometheus Native Histograms Are Now a Stable Feature | Krajo Krajcsovits | 04m |
Prometheus native histograms become a stable feature with a concrete compatibility promise: no breaking API changes until 4.0.0 covering PromQL, JSON results… | ||
| Lightning Talk: Observe everything! Like... for example 3D printers | Pavel Strobl | 05m |
A homemade Prusa exporter written in Go collects printer temperatures, print status and throughput and feeds Prometheus 3.7 for visualization in Grafana. The… | ||
| Lightning Talk: OTel Demo: Showcasing Prometheus and improving its OTel support | Cyrille Le Clerc | 06m |
The OpenTelemetry demo is upgraded to showcase Prometheus integration end to end: it now promotes infrastructure and service resource attributes into… | ||
| Lightning Talk: What if we had PromQL syntax highlighting in MS teams? | Célian Garcia | 02m |
Reverse engineers MS Teams web client, discovers it uses highlightJS, and implements a working PromQL grammar published in a personal GitHub namespace… | ||
| Lightning Talk: Ul To Debug Target Relabeling Rules | Julius Volz | 02m |
Adds a visualizer into the Prometheus UI that shows the discovered labels for each target, then steps through each relabeling rule and highlights diffs (added… | ||
| Lightning Talk: Viewing Raw Samples for Troubleshooting Queries | Roman Khavronenko | 05m |
Prometheus and VictoriaMetrics queries return interpreted results not the stored points, so the precise claim is that seeing raw samples changes… | ||
| Closing / See you next time | Goutham Veeramachaneni | 05m |
Nostalgic wrap up noting the conference returns to Munich and everything mostly goes smoothly, with thanks to sponsors and volunteers and a call to give a… | ||
| nothing matches | ||