Skip to main content

Why did I create SolidPing?

· 4 min read
Florent Clairambault
Maintainer of SolidPing

It starts with a tool I loved, until I didn't.

For years I used a hosted monitoring service for a handful of endpoints. A few HTTP checks, the occasional ping. I only ever cared about one feature: the Slack integration.

One incident, one message in the channel. Every status change after that (failing, slow, recovered, failing again) was threaded under it. No flood, one place to look. The message had an acknowledge button that silenced the pings until the incident actually resolved. Best on-call ergonomics I have used.

Then it moved behind a paywall. Grandfathered for a few months, then pay up or go back to the noisy version. That would have been fair if I needed anything else in that tier. I didn't. I needed that one feature.

So I went looking, and as always I started with open source.

Self-hosting is in a good place​

The list of open source tools I run myself keeps growing: Sentry for errors, Mattermost for chat, OpenObserve for logs and traces, Metabase and Grafana for dashboards, Plausible for analytics, n8n for glue, Vaultwarden for passwords. None of them need a Kubernetes cluster and a DevOps team. They all run on a small VPS.

For monitoring the obvious candidate was Uptime Kuma. It is popular, the integration list is long, it does almost everything. Two things did not click for me. The UI is not to my taste, which is subjective and people who like it are not wrong. And there is no managed version, which mattered to people around me who said "give me a URL and I'll pay you".

I tried others. Open cores with the interesting parts missing, projects untouched for years, one that wanted a Kafka cluster to watch three URLs.

What I wanted​

  1. One binary you drop on a server. Or a container, a compose file, a Helm chart. Running in five minutes.
  2. Linux, Windows and macOS. Plenty of small businesses run Windows and should not have to translate every guide to PowerShell.
  3. Postgres if you have it, SQLite if you don't, same features either way.
  4. Slack alerts done properly. One message per incident, threaded updates, an acknowledge button that actually silences things.
  5. Hostable as a SaaS too, because some people want a managed version and that is how the project pays for itself.

Then Claude Code happened​

I planned something small. HTTP and TCP checks, a status page, that one Slack integration. Six weekends.

I have used Claude Code daily for over a year, and the cost of adding a feature has dropped a lot. Not to zero. Every feature is still a UX decision, a maintenance burden and a docs page. But typing the code stopped being the slow part. So every time I needed something, or an early tester asked for something reasonable, I added it.

Multi-tenant data model. Magic links. Sign in with Slack. Status pages on custom domains. Incidents with timelines. Webhooks. Email digests. Agents that run inside your own network and receive credentials the server cannot decrypt. 40 check types. An llms.txt so the docs are readable by other agents.

That is how the scope got here, and I do not regret any of it. I also know scope creep is how projects become unmaintainable, so the next part is pruning. Keeping it small enough that I can hold the whole thing in my head, and keeping the install boring.

Try it​

If you lost the same Slack integration, or you just want a small self-hostable monitoring tool that will not bury you in YAML, try it and tell me what is missing.

Demo, no signup: solidping.io/demo Source: github.com/fclairamb/solidping