fix: ps cmd should ignore dead runners (#1050)

fixes #1046

Runners that were found to have died (rather than cleanly being killed)
end up with deployment_id & module_name not set to null.
Changes:
- removes all dead runners from showing in the `ps` command
- correctly updates `deployment_id` and `module_name` to null in the
following cases:
    -  Runner terminates unexpectedly
    -  Runner is stale

backend/controller/sql/queries.sql

@@ -110,7 +110,8 @@ RETURNING deployment_id;
 -- name: KillStaleRunners :one
 WITH matches AS (
     UPDATE runners
-        SET state = 'dead'
+        SET state = 'dead',
+        deployment_id = NULL
         WHERE state <> 'dead' AND last_seen < (NOW() AT TIME ZONE 'utc') - sqlc.arg('timeout')::INTERVAL
         RETURNING 1)
 SELECT COUNT(*)
@@ -119,7 +120,8 @@ FROM matches;
 -- name: DeregisterRunner :one
 WITH matches AS (
     UPDATE runners
-        SET state = 'dead'
+        SET state = 'dead',
+            deployment_id = NULL
         WHERE key = $1
         RETURNING 1)
 SELECT COUNT(*)
@@ -160,7 +162,7 @@ SELECT d.min_replicas,
        r.endpoint,
        r.labels AS runner_labels
 FROM deployments d
-         LEFT JOIN runners r on d.id = r.deployment_id
+         LEFT JOIN runners r on d.id = r.deployment_id AND r.state != 'dead'
 WHERE d.min_replicas > 0
 ORDER BY d.name;