Show HN: Ciao – HTTP checks and tests monitoring | Hacker News Reader

Show HN: Ciao – HTTP checks and tests monitoring

brotandgames.com

77 pointsbrotandgames26 comments

Neat. Only looking at the docs and screenshots:

I think you could add a ssl cert expiration warning to this. It's use case is so similar and the data is already there and only needs to be parsed & displayed :)

I use letsmonitor.org for this - it's a free managed service.

We have a lot of internal only web services, so a self hosted option would be a plus.

different usecase. this tool is selfhosted and can resolve internal urls.

Seconded, this would be a killer feature.

Adding TLS cert expiraton warning is one of the things on the roadmap :)

Looks like a very cool solution if you can't use any of the common SaaS solutions out there like Pingdom or UptimeRobot.

Making getting this thing reliable is going to maybe a bit harder. Monitoring systems have this sort of "meta" issue of needing to be more reliable than the things they monitor. I'm building and running one as we speak and I had some interesting cases in that regard.

If you can us a SaaS tool, then here is my obligatory plug of my hosted / SaaS https://checklyhq.com.

> Monitoring systems have this sort of "meta" issue of needing to be more reliable than the things they monitor.

Only slightly; they just need to have an extremely low likelihood of going down at the same time.

It’s okay if your monitoring is down sporadically, just as long as it doesn’t overlap with the device under test also going down unexpectedly.

correct, if you run a dedicated monitoring system for a handful of targets. If you run a SaaS or a very large monitoring system for a large corporation, this doesn't fly anymore. There is simply no moment available to be down. There is always overlap with some system.

Just set up a couple more monitoring servers. You want to test from many regions anyway to detect ISP shenanigans.

To an extent.

A monitoring system that is unreliable risks losing trust. Monitoring systems that aren't trusted tend to be ripped out or ignored.

If it is down too often, people will constantly question whether its silence is an indicator of health.

If it triggers too many false alarms, people will constantly question whether an alarm indicates a real problem. If it gets bad enough, they might start ignoring them.

Could you elaborate on the cases? This would be helpful for all of us.

'Ciao' has an interesting etymology:

https://en.wikipedia.org/wiki/Ciao#Etymology

Apropos for an unattended daemon

We've been using UptimeRobot for this purpose for a number of years. Generally happy. Just something to keep in mind is that we've found their keyword check very handy, and in fact use it more than the status code check. That is, it looks to see if a word (or words) of our choosing are being rendered and pings us know if not. Some of our environments will serve a default 200, but it might not be for the page we hoped for (e.g., a default nginx page after a reboot).

I do like the open source option you've provided though. As I said, we've been using UptimeRobot for years and currently use it for around 50 sites. I fear if/when they change their pricing tiers.

Nice product! And thank you for making it opensource. Any particular reason why you chose Ruby?

A little plug : I have developed https://www.guardscript.com. It is a service that detect any modification made to static files hosted on your website. The goal is to detect any unwanted modification and revert them as soon as possible.

Also check out PageProbe - https://nodetics.com/pageprobe

Looks like at nice replacement for my current solution of a shellscript/cron combo (for monitoring a handful of internal endpoints)

Update: on second thought, without support for webhooks, this is going to be somewhat less useful than my current setup.

Support for webhooks is on the roadmap.

support for webhooks is already there on 1.4.0

Thanks, I'm checking it out atm. Btw., a varname is converted to string on your website: "medium_username@gmail.com"

Looks great. Any reason why the app has on auth despite being based on rails? http basic auth seems extreme.

The idea is to keep the app as simple as possible - without management of users and authentication/authorization methods. This can be accomplished with an auth proxy standing before the app or a similar method.

HTTP Basic auth is considered secure when running over TLS.

I'm using www.websitepulse.com for that stuff

Nice, I like the simplicity around the solution and how easy it is to use.

Along with checklyhq mentioned in the thread if you are looking at comparing SaaS tools in the space check out my tool https://appdoctor.io

Although it does much more then just status checking it may fit someones use case with its advance tests or statuspages.

Reply on news.ycombinator.com