Is TaskId restricted to Ulid by design?

[realonbebeto]

I have been toying with apalis-postgres to push tasks to the backend as follows:

`
let backend: PostgresStorage = PostgresStorage::new(&pool);
let task = TaskBuilder::new(v)
.with_task_id(TaskId::new(tx_id.to_string())).build

backend.push_task(task).await?;`

The following is the error am encountering:
= expected struct apalis::prelude::Task<_, SqlContext<Pool<Postgres>>, Ulid>
found struct apalis::prelude::Task<_, _, std::string::String>

Going by the source code:

apalis/apalis-core/src/task/task_id.rs

Line 87 in 60b0d69

/// Consider using a ulid/uuid/nanoid in backend implementation

the intentions seem to be that it is flexible only that I am not able to figure out how to make use other id types except Ulid.

[geofmureithi]

TaskId implements FromStr you need to use that.
You are passing a String but PG expects a Ulid type.

[realonbebeto]

Well, then implication for that is that TaskId must always be Ulid for PG.

Is it extensible to use an alternative string, uuid, serial?

[geofmureithi]

No, because PG uses the Ulid to order jobs.

…

On Thu, 28 May 2026, 16:05 Bebeto Nyamwamu, ***@***.***> wrote: Well, then implication for that is that TaskId must always be Ulid for PG. Is it extensible to use an alternative string, uuid, serial? — Reply to this email directly, view it on GitHub <#748?email_source=notifications&email_token=AWXVRGW3PKGXMMMOX5M65PT45A2S5A5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZQHEYTCOJZUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-17091199>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWXVRGTQ5DXGEKIM7BVBOVT45A2S5AVCNFSM6AAAAACZPPGX7SVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTOMBZGEYTSOI> . You are receiving this because you commented.Message ID: ***@***.***>

[realonbebeto]

Sure. Thank you.

[geofmureithi]

If you are trying to attach a unique string to a task use idepondency_key

…

On Thu, 28 May 2026, 16:16 Bebeto Nyamwamu, ***@***.***> wrote: Sure. Thank you. — Reply to this email directly, view it on GitHub <#748?email_source=notifications&email_token=AWXVRGRX4TXLFIVRXOGQS6D45A32LA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZQHEYTGNBRUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-17091341>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWXVRGVFLSJFFPDDAHMPH3D45A32LAVCNFSM6AAAAACZPPGX7SVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTOMBZGEZTIMI> . You are receiving this because you commented.Message ID: ***@***.***>

[realonbebeto]

Very much aware of that. Thank you. The reason to think I can have an extensible TaskId is this trait FetchById.

How would you know the TaskId if you dont assign it when pushing the tasks to the backend?

[realonbebeto]

@geofmureithi What is the intended outcome when using the idepondency_key and the same job is submitted twice? Should an error be triggered, an update is made to the job or just ignored?

[geofmureithi]

The approach is usually you generate an Id, use with_task_id. Sinks in rust return a unit. You can also use an sql query if you need to be fine grained

…

On Thu, 28 May 2026 at 17:09, Bebeto Nyamwamu ***@***.***> wrote: Very much aware of that. Thank you. The reason to think I can have an extensible TaskId is this trait FetchById. How would you know the TaskId if you dont assign it when pushing the tasks to the backend? — Reply to this email directly, view it on GitHub <#748?email_source=notifications&email_token=AWXVRGXKJYBJZCIIOFFFV3T45BCC7A5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZQHEZDCOJSUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-17092192>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWXVRGSEMPDNJZV6RGRPEW345BCC7AVCNFSM6AAAAACZPPGX7SVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTOMBZGIYTSMQ> . You are receiving this because you commented.Message ID: ***@***.***>

[realonbebeto]

Thats the correct 👌 approach with using with_task_id but then as you mentioned there is a design constraint of Ulid (thinking out loud is if this can be flexible to other id types so that if someone is using uuids they can use that instead of having an extra dependency) by PG to order jobs.

[geofmureithi]

You should be able to convert Ulid to Uuid

…

On Thu, 28 May 2026, 17:22 Bebeto Nyamwamu, ***@***.***> wrote: Thats the correct 👌 approach with using with_task_id but then as you mentioned there is a design constraint of Ulid (thinking out loud is if this can be flexible to other id types so that if someone is using uuids they can use that instead of having an extra dependency) by PG to order jobs. — Reply to this email directly, view it on GitHub <#748?email_source=notifications&email_token=AWXVRGQGEJO77VJAWEQ3RKD45BDRNA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZQHEZDIMJXUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-17092417>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWXVRGUMKCTI6JT53QHAFBL45BDRNAVCNFSM6AAAAACZPPGX7SVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTOMBZGI2DCNY> . You are receiving this because you commented.Message ID: ***@***.***>

[realonbebeto]

Thats the correct position. I am curious how that can be achieved with having uuid only in the dependencies. I posit that both uuid and ulid have to be dependencies in the project which beats the concern of having only one instead of both.

[geofmureithi]

You need ulid with the uuid feature plus uuid in deps

[realonbebeto]

As its stands thats the only way to do it.

[geofmureithi]

It should throw a conflict error. https://docs.rs/apalis-postgres/latest/src/apalis_postgres/sink.rs.html#137

…

On Sat, 6 Jun 2026 at 04:24, Bebeto Nyamwamu ***@***.***> wrote: @geofmureithi <https://github.com/geofmureithi> What is the intended outcome when using the idepondency_key and the same job is submitted twice? Should an error be triggered, an update is made to the job or just ignored? — Reply to this email directly, view it on GitHub <#748?email_source=notifications&email_token=AWXVRGRSR62JMQDJCEW2HVT46NXFPA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZRHE4TCOJZUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-17199199>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWXVRGUULKUFYMLAEUMPFN346NXFPAVCNFSM6AAAAACZPPGX7SVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTOMJZHEYTSOI> . You are receiving this because you were mentioned.Message ID: ***@***.***>