Prometheus
Open-source time-series monitoring and alerting system that scrapes and stores metrics.
HN commenters use it as the metrics layer of a self-hosted Grafana stack, often alongside Loki,1234 and for alerting on failed systemd services and timers.56 The most common complaint is that scaling it horizontally gets complicated,78 and some find it too heavy for small devices or simple setups.910
Why people recommend it
Grafana stack · 3
Commonly run with Grafana, Loki and related tools as a complete open-source monitoring stack.
“just using simple scripts + graphana + prometheus + Loki and for alloy/nodeexporter”
Show 1 more
“a full monitoring stack (Grafana + Prometheus + Loki + Grafana Alloy, etc.)”
Self-hosted monitoring · 3
A regular part of self-hosted and homelab setups for monitoring services.
“Grafana + Prometheus monitoring everything.”
“grafana/prometheus”
Show 1 more
“Prometheus, grafana & monitorix: monitoring”
Service alerting · 3
Used to monitor systemd timers and services and to alert on failures or unusual activity.
“I use systemd timers, Prometheus and alert on failed services”
“I use systemd timers which are monitored through prometheus.”
Show 1 more
“using things like Grafana+Prometheus to monitor your personal computer/workstation as well. This is what I've done”
Remote write · 1
Although pull-based, it can remote-write metrics to external storage.
“Even Prometheus, while pull-based inside the cluster, can remote-write to external storage”
Caveats and dissent
Horizontal scaling · 2
Commenters say scaling is more complicated and a better horizontally scalable option is needed.
“Prom needs a better horizontally scalable solution.”
Resource weight · 1
One commenter found Prometheus and Grafana too heavy for small devices.
“Monitoring (Prometheus/Grafana) was too heavy for small devices.”
Setup overhead · 1
One user says the setup may be too much for many simple setups.
“I use systemd timers, Prometheus and alert on failed services”
Multiple query languages · 1
In the Grafana stack, each data source has its own query language, unlike Datadog.
Often compared with
Who recommends it in their own words
From the makers not counted in the ranking
“Observability stack — Prometheus, Grafana, dashboards, alerting, and on-call workflows”
Threads where it came up
- What I self-host 1 mention
- Ask HN: What do you run on a $5 VPS that's worth keeping online 24/7? 1 mention
- Show HN: Sigabrt.dev – cronjob monitor with an SSH TUI 1 mention
- Feeling Sad about AI 1 mention
- Ask HN: Freelancer? Seeking freelancer? (September 2026) 1 mention
- OTel isn’t going well 1 mention
Data: prometheus.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 15
“grafana/prometheus”
“Prometheus, grafana & monitorix: monitoring”
“I use systemd timers, Prometheus and alert on failed services”
“installed Prometheus and Grafana and built grafana dashboards all using AI”
“Prometheus and Grafana”
“OSS LGTM/P works good for me.”
“Even Prometheus, while pull-based inside the cluster, can remote-write to external storage”
“I use systemd timers which are monitored through prometheus.”
“using things like Grafana+Prometheus to monitor your personal computer/workstation as well. This is what I've done”
“a full monitoring stack (Grafana + Prometheus + Loki + Grafana Alloy, etc.)”
“Give Claude/Codex access to observability (Grafana, Prometheus, Tempo, PagerDuty)”
“Grafana + Prometheus monitoring everything.”
“just using simple scripts + graphana + prometheus + Loki and for alloy/nodeexporter”
“Monitoring (Prometheus/Grafana) was too heavy for small devices.”
“Prom needs a better horizontally scalable solution.”