Bug 1891815 - Increase global task timeout to two hours

We have bitrisescript tasks that take longer than an hour and it looks
like we currently have no way to set a per script timeout (see the
comment about json-e in scriptworker.yml).

So the quick fix here is to bump the global timeout. I believe it's
extremely rare (if not impossible) for us to hit this timeout anyway,
because Kubernetes has its own timeout (`terminationGracePeriodSeconds`)
that is much shorter than this one. The only reason I even knew this
timeout exists is because I had set the above to `7200` for
bitrisescript. All other scripts have this value set to `3600` or less,
which means they should get force killed by k8s long before we ever
reach the scriptworker timeout.

docker.d/init.sh

@@ -113,7 +113,7 @@ if [ "$ENV" == "prod" ]; then
 fi
 export TASK_CONFIG=$CONFIG_DIR/worker.json
 export TASK_LOGS_DIR=$ARTIFACTS_DIR/public/logs
-export TASK_MAX_TIMEOUT=3600
+export TASK_MAX_TIMEOUT=7200
 export TASK_SCRIPT=$APP_DIR/bin/${PROJECT_NAME}script
 export TEMPLATE_DIR=$APP_DIR/docker.d
 export VERBOSE=true

docker.d/scriptworker.yml

@@ -10,7 +10,7 @@ credentials:
 # artifact_upload_timeout: { "$eval": "ARTIFACT_UPLOAD_TIMEOUT" }
 # task_max_timeout: { "$eval": "TASK_MAX_TIMEOUT" }
 artifact_upload_timeout: 1200
-task_max_timeout: 3600
+task_max_timeout: 7200
 task_script:
     - { "$eval": "TASK_SCRIPT" }
     - { "$eval": "TASK_CONFIG" }