The following lesson content assumes that you have read access to the FME Flow database, and publish permissions on ArcGIS Server and ArcGIS Monitor.
Warning: The rest of this article covers accessing the FME Flow database directly, whether it’s PostgreSQL or SQL Server. This is risky territory (and probably not supported). You don’t want to edit the FME Flow tables, so proceed with caution and use a separate read-only account against the FME schema whenever possible. Some of the queries also need to exist as views in the FME Flow database to work as query layers in ArcGIS, so you’ll need create-view permissions as well.
These queries/dashboards have been tested on:
- An FME Flow 2025.1 instance using a PostgreSQL 16 database, paired with ArcGIS Enterprise 11.5
- An FME Flow 2026.2 instance using a SQL Server 2025 database, paired with ArcGIS Enterprise 12.1
- ArcGIS Monitor 2026.0 (the ability to execute database queries was added in Monitor 2026)
We’ll be working with the following tables in the FME Flow database. In a default installation, FME Flow installs its own version of PostgreSQL. It can also be configured to work with SQL Server and Oracle.
Job History & Status:
- fme_jobs: Data about jobs that are currently running or queued.
- fme_job_history: Information about jobs that have been cancelled, failed, or completed.
- fme_job_status: A lookup table for the status IDs used in the job tables.
Nodes & Engines:
- fme_engines: Information about the engines that have registered into the server pool.
- fme_core_node: Information about the cores that have been installed and are connected to the cluster.
Now that we have the tables described, let’s go over the queries that we are going to be using.
The example queries below are for a PostgreSQL database. SQL Server versions can be found in our GitHub repository.
All Jobs Last 48 Hours:
This query combines currently running jobs with job history to create a single view for use in the ArcGIS Dashboard. It selects jobs completed, cancelled, or failed within the last 48 hours. A 48-hour window was chosen because it’s enough time to spot a trend without loading the dashboard with more data than it needs.
SELECT *
FROM (SELECT jh.job_id,
jh.date_queued,
jh.date_started,
jh.date_finished,
(jh.elapsedtime / 1000)::bigint AS runtime,
jh.tag,
jh.repository_name,
jh.workspace_name,
CASE --clean up the machine names
WHEN jh.engine_instance LIKE '%.%'
THEN split_part(jh.engine_instance, '.', 1)
ELSE split_part(jh.engine_instance, '_', 1)
END AS engine_instance,
js.job_status AS status,
CASE --Create a field to aid with sorting running and queued jobs.
WHEN date_started IS NULL AND date_finished IS NULL THEN 1 -- queued
WHEN date_started IS NOT NULL AND date_finished IS NULL THEN 2 -- running
ELSE 3 -- canceled, finished, etc
END AS sorder,
ua.username
FROM public.fme_job_history AS jh
INNER JOIN public.fme_job_status AS js
ON jh.job_status = js.id
INNER JOIN public.fme_useraccount AS ua
ON jh.useraccount_id = ua.useraccount_id
WHERE jh.date_finished >= NOW() - INTERVAL '48 hours'
UNION ALL
SELECT j.job_id,
j.date_queued,
j.date_started,
j.date_finished,
(j.elapsedtime / 1000)::bigint AS runtime,
j.tag,
j.repository_name,
j.workspace_name,
CASE --clean up the machine names
WHEN j.engine_instance LIKE '%.%'
THEN split_part(j.engine_instance, '.', 1)
ELSE split_part(j.engine_instance, '_', 1)
END AS engine_instance,
js.job_status AS status,
CASE --Create a field to aid with sorting running and queued jobs.
WHEN date_started IS NULL AND date_finished IS NULL THEN 1 -- queued
WHEN date_started IS NOT NULL AND date_finished IS NULL THEN 2 -- running
ELSE 3 -- canceled, finished, etc
END AS sorder,
ua.username
FROM public.fme_jobs AS j
INNER JOIN public.fme_job_status AS js
ON j.job_status = js.id
INNER JOIN public.fme_useraccount AS ua
ON j.useraccount_id = ua.useraccount_id) AS all_jobs
ORDER BY sorder, date_queued DESC;
Save this query as a view in your FME Flow database and provide your ArcGIS Server account read access to the view.
FME Nodes & Engines:
This query pulls the nodes and engines currently registered in the FME Flow database and combines them into a single view. It answers the earlier question: are all the engines registered in the cluster? In a high-availability deployment, this is how you’d catch an Engine (or even a Core) that’s silently dropped out.
SELECT row_number() OVER (ORDER BY name) AS id,
name,
hostname,
engine_manager_node_name,
role,
job_id,
is_leader,
numstandardengines
FROM (SELECT
CASE
WHEN fme_engines.name ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_engines.name
WHEN fme_engines.name LIKE '%.%'
THEN split_part(fme_engines.name, '.', 1)::character varying
ELSE split_part(fme_engines.name, '_', 1)::character varying
END AS name,
CASE
WHEN fme_engines.hostname ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_engines.hostname
WHEN fme_engines.hostname LIKE '%.%'
THEN split_part(fme_engines.hostname, '.', 1)::character varying
ELSE split_part(fme_engines.hostname, '_', 1)::character varying
END AS hostname,
CASE
WHEN fme_engines.engine_manager_node_name ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_engines.engine_manager_node_name
WHEN fme_engines.engine_manager_node_name LIKE '%.%'
THEN split_part(fme_engines.engine_manager_node_name, '.', 1)::character varying
ELSE split_part(fme_engines.engine_manager_node_name, '_', 1)::character varying
END AS engine_manager_node_name,
'Engine'::text AS role,
fme_engines.job_id,
NULL::boolean AS is_leader,
NULL::integer AS numstandardengines
FROM fme_engines
WHERE fme_engines.type <> 2
UNION ALL
SELECT
CASE
WHEN fme_core_node.name ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_core_node.name
WHEN fme_core_node.name LIKE '%.%'
THEN split_part(fme_core_node.name, '.', 1)::character varying
ELSE split_part(fme_core_node.name, '_', 1)::character varying
END AS name,
CASE
WHEN fme_core_node.name ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_core_node.name
WHEN fme_core_node.name LIKE '%.%'
THEN split_part(fme_core_node.name, '.', 1)::character varying
ELSE split_part(fme_core_node.name, '_', 1)::character varying
END AS hostname,
'NA'::character varying AS engine_manager_node_name,
'Core'::text AS role,
NULL::integer AS job_id,
fme_core_node.is_leader,
fme_core_node.numstandardengines
FROM fme_core_node
UNION ALL
SELECT
CASE
WHEN fme_remote_engine_conn.name ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_remote_engine_conn.name
WHEN fme_remote_engine_conn.name LIKE '%.%'
THEN split_part(fme_remote_engine_conn.name, '.', 1)::character varying
ELSE split_part(fme_remote_engine_conn.name, '_', 1)::character varying
END AS name,
CASE
WHEN fme_remote_engine_conn.name ~ '^\d+\.\d+\.\d+\.\d+$'
THEN fme_remote_engine_conn.name
WHEN fme_remote_engine_conn.name LIKE '%.%'
THEN split_part(fme_remote_engine_conn.name, '.', 1)::character varying
ELSE split_part(fme_remote_engine_conn.name, '_', 1)::character varying
END AS hostname,
'NA'::character varying AS engine_manager_node_name,
'Remote Engine'::text AS role,
NULL::integer AS job_id,
NULL::boolean AS is_leader,
fme_remote_engine_conn.numstandardengines
FROM fme_remote_engine_conn) t;
Save this query as a view in your FME Flow database and provide your ArcGIS Server account read access to the view.