ariejan/apocalypse

Warns you about the coming server apocalypse

apocalypse's Issues

Alarm notification: SMS

Get an SMS when an alarm is triggered.

Send apocalypse-client version and notify if outdated

There may be updates to apocalypse-client that go unnoticed. By sending the AC version with every metrics update we can warn about this in the dashboard.

Allow override of server name

Now we use the host_id. This is okay in most cases, but EC2 instances look ugly. It'd be nice if we could override this to a more elegant/readable name.

Make last updated times shorter

E.g. '24s ago' instead of 'about 24 seconds ago'. This makes everything fit more neatly on a screen.

Unknown fields are to prominent, being bright yellow.

Unknown fields are to prominent, being bright yellow.

Run capybara scenarios against a remote app (web only)

Allow a capybara scenario to be executed agains a host. This works like any normal test and should give a positive (all assertions passed) or negative (it didn't work) result.

This would be an awesome feature to check if an application is actually working or not.

Define Data API

Define the API that allows the GUI and third-party apps to access monitoring data from Apocalypse.

No per-host configuration possible

It should be easy enough to insert some values into redis to set thresholds for each host. This is important, especially for multi-core system etc.

Last updated gets stale in dashboard

When an update comes in a relative time is shown (e.g. 3 seconds ago). This information is now only updated when a new update comes in. However, after any period of time, the status still says "3 seconds ago".

By setting some HTML-5 property we should be able to have some JS calculate the time_ago each few seconds.

New hosts don't get added to the dashboard dynamically.

When a new hosts sends its first metrics package, the dashboard should be updated to add that host.

Define monitoring API

Define the monitoring API - how do servers communicate data and statistics to the Apocalypse server?

Disk usage always shows 'unknown' in the background

@avdgaag could you look into that?

Second precision for updated_at

Last updated is now using minutes as most significant time indicator. We should use seconds, because that's how ofter we update our stats.

Keep stats for a longer period of time to generate graphs

We could keep data for a longer period of time (depending on redis capacity) and generate sexy graphs from that.

Alarm notification: email

Notify of alarms to specified email addresses.

Ubuntu 11.04 does not return disk info

The current apocalypse-client gem does not appear to repost disk usage correctly.

Check reachability

According to the current architecture there will be a scheduler available that will ping each host at a certain interval. This ping is short (e.g 2-4 messages) and should indicate if the host is up and responding.

This could an 'is up' icon on the dashboard.

Write proper (and up-to-date) installation instructions

The installation docs are a bit out-dated, update them.

Last update shows "-1 minute"

This is due to time difference between server and client. In this case the server is posting times into the future (according to the client).

Display (static) host information

Apocalypse-client sends more information that we currently need, like the number of CPU cores, available RAM and swap etc. It would be great to have this information available somewhere on the dashboard.

Add security

Authentication key per server?

How would we distribute these keys?