Case study · Financial services · Datadog LaunchPad
Dashboards and alerting a pensions group's teams can run themselves
A UK pensions group is turning its Datadog rollout into dashboards and alerting its lean teams can run, and found the cause of a slowdown straight away.
At a glance
Straight away
to the cause of a slowdown
Over 90%
of the main Azure estate covered by Datadog
30/60/90 day
plan agreed with the team
PensionsUKAzureDatadogDatadog LaunchPad
01
The challenge
The challenge
Getting data into an observability platform is the quick part. Turning it into views each team trusts is harder: dashboards and alerts only work when every resource is tagged the same way. The group had already rolled Datadog out across a large Azure estate, and chose specialists to make that data something its lean teams could run themselves, ready for the businesses joining the group.
02
What we did
What we did
- Ran a Datadog HealthScan and agreed a 30, 60 and 90 day improvement plan with the team.
- Worked with the infrastructure team to standardise how Datadog is deployed across Azure.
- Built dashboards with the people who use them, for each core pensions platform, the wider estate and a new AI benefits product.
- Agreed a single tagging standard and designed alert routing, managed as code, so each alert reaches the team that owns the system.
- Agreed cost controls, such as switching off monitoring for idle non-production environments.
03
The outcome
The outcome
- When users reported a slow pensions application, the team saw straight away which process was slowing the affected server, helped by Datadog's AI assistant.
- Datadog deployment standardised across over 90% of the main Azure estate, with every exception accounted for.
- Dashboards shaped around each team, a clear model for routing every alert, and a way to keep Datadog spend in step with use.
This case study is anonymised, so the customer is not named.
Talk to us about Datadog LaunchPad.
Tell us what you run today and what you need from it. We’ll recommend the simplest first step.