aheart / hearth Goto Github PK

Integrate with coveralls or codecov.
Add badge to README.md

something like this should would improve UX for using multiple config files:

./hearth --config my_config.toml

https://github.com/clap-rs/clap is a good candidate.

main.rs is cluttered and should be cleaned up.

Hello aheart,
I wonder if it's possible to implement a ssh jumphost feature (in ssh config as ProxyJump) to connect through one ssh host to other hosts which are not directly accessible.
regards
dot

At the moment it's in Beta, but it might be a good idea to start testing now.

Add online status boolean to https://github.com/aheart/hearth/blob/master/src/metrics/aggregator.rs#L15
derive(Default) sets booleans to false so it's perfect for us.
On the frontend show {{up}}/{{total}} instead of just total.
Where up is the count of servers with online status boolean set to true.

To make it easier to see disk bottlenecks different colors can be used to represent the utilization similarly to Load Average.

iostat can examined to find out how disk utilization is computed.
https://github.com/sysstat/sysstat/blob/master/iostat.c

• Add CHANGELOG
• Remove note about linux kernels < 3.14
• List new features in readme, preferably with gifs

Example from HTOP

Currently aggregated metrics are only displayed as text at the very top of the page.
Charts would be very helpful to visualize a bit of history.

Currently sda is monitored by default

Add screenshot, description and setup guide.

Example 75a8ecb

There are no tests in the repository at the moment. In order to setup Travis CI at least one test is needed.

https://github.com/actix/actix-web/blob/master/CHANGES.md

Retrieving uptime and the number of CPUs should be moved out of SshClient
https://github.com/aheart/hearth/blob/master/src/ssh.rs#L12-L13

The number of cpus can be extracted from /proc/stat

Currently metrics are aggregated in a vector, which is limited to 11*120.

htop can be used as a source of inspiration

Currently data is fetched after 1 second + the time it takes to fetch the data. Which can result in inconsistent intervals in scope of one server and across multiple servers

Sparklines is abandoned and missing critical features.
New Charts should be able to:

• work with both X and Y axis data (support time/date)
• display stacked data
• render very quickly
• have a license that allows usage in open source projects for free

Example 75a8ecb
SystemTime should probably be injected into DiskStats from without

Line charts are not very convenient in this case because they keep overlapping and transparency doesn't help.

Current timeout is too large and should be set to 1 second.

Currently it's only possible to authenticate using the ssh agent.
Private key authentication would be very helpful for cases when you want to run hearth in the background.

https://docs.rs/ssh2/0.3.1/ssh2/struct.Session.html#method.userauth_pubkey_file

Metrics should be aggregated into separate buffers to accommodate all the timeframes.
The UI should provide the option to twitch between 1, 5, 15 second intervals.
The setting should be stored in the WS session.
When changing an interval a WS message should be send to the WS server.
After that a batch of metrics should be resent to re-render all the charts.

Example 75a8ecb

Example:

$ cat /sys/class/net/eth0/statistics/rx_bytes /sys/class/net/eth0/statistics/tx_bytes
481853516786
903686511858

Network interface must be configurable

Integration with Travis is done.
.travis.yml is required with instructions to build the code and run tests.

A new release is coming soon. First alpha has been released.

I have found that Hearth works on OSX just great:

It would be good to have executable for Mac. It might be automated using Travis:
https://docs.travis-ci.com/user/reference/osx/