Recurring Cleanup with Profiles

Use this workflow when:

you already know the cleanup shape you want
you are tired of rerunning a long cleanup command
you want a stable preset you can adjust over time

Do not start here if you have not yet done a narrow dry-run manually.
First validate the command shape with First Run or Rolling Retention.

Step 1: inspect available profile fields

dmd profile fields

This shows the supported stored field names and the value shapes accepted by --set.

Step 2: create a baseline profile

Example:

dmd profile add nightly-dms \
  --set include_ids=<id> \
  --set keep_within=2w \
  --set keep_last=20 \
  --set preserve_cache=true

This creates a baseline profile that can seed the preserve-cache before you introduce a narrow fetch window.

Replace <id> with a guild, category, guild text channel, DM, or Group DM ID. If you want multiple IDs, use a comma-separated list such as include_ids=123,456.

The scope field is not required by the CLI. If you omit include_ids and exclude_ids, the profile uses the default cleanup scope: all eligible channels.

Step 3: inspect the stored result

dmd profile show nightly-dms

Do this immediately after creation so you can confirm:

the right fields were stored
nothing unexpected ended up in the profile
the shape is what you want to maintain over time

Step 4: preview the profile as a dry-run

dmd clean --profile nightly-dms --dry-run

This is the most important step.

You want to verify:

the selected scope is right
the retention behavior looks right
the number of affected messages and reactions is believable

Step 5: seed the real run once

dmd clean --profile nightly-dms

This first real run intentionally has no fetch_within limit. It lets the tool inspect the full scoped history and populate preserve-cache for kept messages.

Step 6: add the recurring fetch window

After the first real run has seeded preserve-cache, update the profile for the recurring schedule:

dmd profile update nightly-dms --set fetch_within=1d

For a daily run, 1d means the next run fetches the interval since the previous run while preserve-cache supplies older kept messages.

Use a wider fetch window if you want each recurring run to fetch the whole keep window too:

dmd profile update nightly-dms --set fetch_within=2w1d

Step 7: evolve the profile instead of rewriting commands

Examples:

dmd profile update nightly-dms --set verbose=1
dmd profile update nightly-dms --set max_messages=none
dmd profile update nightly-dms --unset fetch_within

Use this when you want to:

widen or narrow retention
change output defaults
remove temporary limits

When preserve-cache matters here

If your recurring profile combines:

keep_last
limited fetch_within
repeated runs

then preserve_cache usually becomes worth enabling.

That is the combination most likely to need continuity between runs.

See Preserve Cache for the detailed behavior.

Common mistakes

Creating a profile before validating the command manually.
Creating a profile without an explicit include_ids or exclude_ids scope.
Forgetting to run the saved profile once in --dry-run.
Storing too many output preferences before you know what level you actually want.
Using keep_last with a limited fetch window but no preserve-cache.