I've inherited an Azure project and am new to the platform. It appears to have a PostgreSQL database, a couple of VMs, scheduled jobs, and a web app, but I'm struggling to see what's running, whether it's healthy, and where the roughly $170 monthly bill comes from.
The company handles sensitive information. Where should I start?
Start with an inventory before changing anything. In the subscription, look through the resource groups and list the resources, their types, locations, and owners if those are documented. Check whether you have access to every subscription involved, since one subscription won't necessarily show the whole environment.
Then trace the app outward. Find where the web app is hosted, what runs the scheduled jobs, how they reach PostgreSQL, and what the VMs are for. Ask for the deployment instructions and any existing architecture notes. A simple diagram and a list of unknowns will make the monitoring work much easier.
For health monitoring, look at Azure Monitor and Application Insights. Check which resources already send metrics and logs, then fill the gaps. Some guest-level VM data needs a monitoring agent, so don't assume the portal has everything just because the VM is listed.
Set alerts around things you actually need to act on, such as failed scheduled jobs, application errors, and resource availability. A job can exist and still fail every run, so make sure its success and failure are visible somewhere.
Use Cost Management to break the bill down by resource and service. Set a budget alert too. The monthly total alone won't tell you whether the database, VMs, storage, or something else is driving the spend.
Before stopping or resizing anything, find out what depends on it and whether there are backup or retention requirements.
The portal is useful for exploring, but it doesn't have to be the long-term source of truth. Once you understand the setup, consider describing the infrastructure in Bicep or Terraform, keeping it in a repository, and deploying changes through a pipeline.
Don't rush to redeploy the inherited environment just to make it neat. First document what's there and work out how to bring existing resources under management without disrupting them.
Azure may or may not be the right amount of infrastructure for this app. Before switching stacks, though, get a clear picture of the security requirements and what the existing services actually do. A cheaper or simpler setup still needs access controls, backups, and monitoring.
Answers are reviewed before publication.
x640!clasp_2:
Also check who has access and how deployments authenticate. Sensitive data is a good reason to sort out ownership before handing over shared credentials.