pygeoapi is a Python server implementation of the OGC API suite of standards. The project emerged as part of the next generation OGC API efforts in 2018 and provides the capability for organizations to deploy a RESTful OGC API endpoint using OpenAPI, GeoJSON, and HTML. pygeoapi is open source and released under an MIT license.
This makes the base process manager add the job to it's database before returning any response to the client.
Related Issue / discussion
In my testing, I send many async jobs requests at the same time. Once the response header comes back, I ping the server every second to see if the job succeeded.
On my slow VM, with more than 10 simultaneous requests, the first ping fails because the job doesn't exist. I receive the following response from the server : {'code': 'InvalidParameterValue', 'type': 'InvalidParameterValue', 'description': '<job id>'}.
Diving in the code, I see that the first add_job call of the manager, that registers the job to the db, happens within the execution function. In the async case, this means that the Process creation and start happens between the initial response and the add_job . In the case of (relatively) large loads like in my example, this takes more than 1 sec.
In any case, it feels better if the response sent with the accepted status only happens once the job is registered to the database.
Additional information
I just started playing with pygeoapi, I might have missed something that justified the previous behaviour!
Dependency policy (RFC2)
[x] I have ensured that this PR meets RFC2 requirements
[x] I'd like to contribute this bugfix to pygeoapi. I confirm that my contributions to pygeoapi will be compatible with the pygeoapi license guidelines at the time of contribution
[x] I have already previously agreed to the pygeoapi Contributions and Licensing Guidelines
Hi!
Overview
This makes the base process manager add the job to it's database before returning any response to the client.
Related Issue / discussion
In my testing, I send many async jobs requests at the same time. Once the response header comes back, I ping the server every second to see if the job succeeded.
On my slow VM, with more than 10 simultaneous requests, the first ping fails because the job doesn't exist. I receive the following response from the server :
{'code': 'InvalidParameterValue', 'type': 'InvalidParameterValue', 'description': '<job id>'}.Diving in the code, I see that the first
add_jobcall of the manager, that registers the job to the db, happens within the execution function. In the async case, this means that the Process creation and start happens between the initial response and theadd_job. In the case of (relatively) large loads like in my example, this takes more than 1 sec.In any case, it feels better if the response sent with the
acceptedstatus only happens once the job is registered to the database.Additional information
I just started playing with pygeoapi, I might have missed something that justified the previous behaviour!
Dependency policy (RFC2)
Updates to public demo
Contributions and licensing
(as per https://github.com/geopython/pygeoapi/blob/master/CONTRIBUTING.md#contributions-and-licensing)