opsweekly

Opsweekly

What is Opsweekly?

Opsweekly is a weekly report tracker, an on call categorisation and reporting tool, a sleep tracker, a meeting organiser and a coffee maker all in one.

The goal of Opsweekly is to both organise your team into one central place, but also helps you understand and improve your on call rotations through the use of a simple on call "survey", and reporting as a result of that tracking.

Alert classification is a complicated task, but with Opsweekly a few simple questions about each alert received can pay dividends in improving the on call experience for your engineers.

Features

• Weekly Updates: Every member of your team can write a weekly status update using hints (e.g. Github commits, JIRA tickets) to inform the team what they've been working on, and then optionally email it out.
• On-Call Alert Classification: Track, measure and improve your on call rotations by allowing your engineers to easily classify and document the alerts they received.
  • Make a simple assesment for each alert that relates to whether action was taken, no action was taken, or whether the alert needs modification for follow up later
  • Free notes field to allow documentation of actions taken to refer back to later
  • Bulk classification for time saving
• Sleep Tracking: If your engineers have popular life tracking devices such as Fitbit or Jawbone UP, they integrate with Opsweekly to provide even more insight into the effect on call is having on their lives.
  • Mean time to sleep (MTTS) and sleep time lost to notifications are calculated.
  • Easy to configure but gives valuable data that could lead to questions like "can this alert wait until morning as it keeps waking up our engineers?"
• In depth reporting: As you start to build up data, Opsweekly starts to generate reports and graphs illustrating your on call rotations.
  • Examples include: action taken vs no action taken on alerts, what alerts wake people up the most, mean time to sleep, top notifying hosts/services, average alert volume per day, and how on call has improved (or not) over the last year
• Personal reporting: As well as a summary for all rotations, users are able to gain insight into their own behaviours.
  • How have their on calls affected them?
  • How much sleep do they lose on average?
  • How does this compare to others?
  • Optionally, one can view a sleep retrospective that compares the last several on-call rotations' impact on sleep loss.
    • In phplib/config.php, define oncall_sleep_retrospective_count with a numeric value (such as 3). Users viewing their profile will then see how past weeks affected their sleep.
• Meeting Mode: Make running a weekly meeting simple with all the data you need in one page, and a facility for people to take notes.
  • Meeting mode hides all UI displaying only information required for the meeting.
  • The on call report for the previous week is included, along with key stats and elements from report
  • All weekly updates are displayed in case items need to be discussed
  • Set up a cron to remind people about the weekly meeting and provide the permalink to the meeting
• Powerful Search: All data is searchable using a powerful search function. The default search mode is fuzzy, which will return results from all data stored in Opsweekly. However, you can get more specific:
  • Search previous on call alerts for a history of that alert, previous engineer's notes, how the alerts were classified (is this alert constantly "no action taken?") and a time map showing it's frequency over the past year.
  • Search Weekly Updates for full context on changes made previously
  • Search Meeting Notes for agenda items discussed in previous meetings
• Fully timezone aware: Obviously it's important for users to be editing the alerts they receive in the timezone that they received them in. Each user can set their own timezone for the whole Opsweekly UI.
• Fill in as you go/drafts: Both the Weekly report and the On-call reports can be updated to multiple times during the week, so the user does not have to edit a hefty report at the end.

Screenshots

Please visit the screenshot README for a guided tour of how Opsweekly works and the reports it can generate!

Prerequisites

• A webserver
• PHP 5.4 (or higher), including the curl extensions for PHP, MySQL extension, and short_open_tags enabled
• MySQL for data storage

Installation/configuration

1. Download/clone the repo into an appropriate folder either in your webservers directory or symlinked to it. or:
2. Create a configuration in your webserver for Opsweekly, if using it as a seperate domain (e.g. VirtualHost)
3. You must increase the PHP variable max_input_vars for submitting on-call reports. See Increasing max input vars
4. Create a MySQL database for opsweekly, and optionally grant a new user access to it. E.g.:
   • mysql> create database opsweekly;
   • mysql> grant all on opsweekly.* to opsweekly_user@localhost IDENTIFIED BY 'my_password';
5. Load the database schema into MySQL, e.g. mysql -u opsweekly_user opsweekly

e.g., using cron, weekly at 2pm: 0 14 * * 3 php /var/www/opsweekly/send_meeting_reminder.php myweekly.yourdomain.com

Known issues/caveats/future goals

• As the name implies, Opsweekly is rather tied to the concept of a week. In theory the database stores time ranges, but the UI is all based on a week's worth of data
  • At some point I invisage dropping the concept of a fixed time period and instead having "providers" that pull the periods people were on call, prompting them to fill in the data. E.g. Pagerduty: You were on call from X to Y, please categorise your alerts for that period.
• Whilst users can fill in their weekly report and on call report as the week continues, two people cannot edit the same on call report otherwise duplicate events may appear in the reports. This is due to the timestamp being part of the unique key for alerts.
• Garbage in, garbage out: There is no way to exclude items returned by the on call providers in the reports right now, outside of deleting anything from the database you're not happy with after compiling the on-call report. I want to add a "soft delete" so deliberately allow editing of reports for unforseen reasons (e.g. monitoring system goes wrong and didn't actually send any alerts but they were logged)