One of the feature of Situation is that agents can collaborate once they share the same database (currently, only postgres is supported).
Database
The first step is to provision a PostgreSQL database.
You can choose to host an instance on-premise (with docker for instance) or let a remote service do it for you (like supabase or any cloud provider).
Here is an example with docker:
docker run \
--rm \
--detach \
--name situation-pg \
-e "POSTGRES_PASSWORD=password" \
-e "POSTGRES_USER=user" \
-e "POSTGRES_DB=situation" \
-p "5432:5432" \
--health-cmd="pg_isready -U user" \
--health-interval=2s \
--health-timeout=2s \
--health-retries=2 \
postgres:17.6
The paramount thing is to get the DSN of the instance so as to pass it to the agents. In the example above it is something like postgres://user:password@[ENDPOINT]/situation?sslmode=disable.
Agents
Once your db is ready, you can deploy agents anywhere on your infra. We advise to do it incrementally since IT mapping generally means IT exploring.
Rule of thumbs
- deploy one agent per subnet since every agent is able to grab data from neighboring hosts
- deploy agent on hosts where exhaustive data collection is mandatory
On each host where agent is deployed, follow the next steps.
Identifier
Every agent has an internal ID that must be unique across your swarm. Stock agent has a default identifier that must be changed.
situation refresh-id
situation.exe refresh-id
Important:
Once the agent has its unique ID, it must keep it and stay on the same host.
You can check the new id by calling the id subcommand.
Execution
situation run --db="postgres://user:password@[ENDPOINT]/situation?sslmode=disable"
situation.exe run --db="postgres://user:password@[ENDPOINT]/situation?sslmode=disable"