Creating a data clip fails on indexing when it's too large

[sentry-io]

Sentry Issue: LIGHTNING-GK

Postgrex.Error: ERROR 54000 (program_limit_exceeded) string is too long for tsvector (1621086 bytes, max 1048575 bytes)
  File "lib/ecto/adapters/sql.ex", line 1096, in Ecto.Adapters.SQL.raise_sql_call_error/1
  File "lib/ecto/repo/schema.ex", line 834, in Ecto.Repo.Schema.apply/4
  File "lib/ecto/repo/schema.ex", line 415, in anonymous fn/15 in Ecto.Repo.Schema.do_insert/4
  File "lib/lightning/runs/handlers.ex", line 326, in anonymous fn/2 in Lightning.Runs.Handlers.CompleteStep.update_step/2
  File "lib/lightning/repo.ex", line 31, in anonymous fn/1 in Lightning.Repo.transact/2
...
(3 additional frame(s) were not displayed)

[taylordowns2000]

@josephjclark , this is what's causing the elixir error in https://github.com/OpenFn/thunderbolt/issues/462 and #2943 (seemingly not DB connection/timeout.) we are prioritizing it now.

cc @theroinaochieng

Note that we must add helptext to the SEARCH bar on history explaining that we cannot search very large dataclips. (Need human friendly language about how large, and what to do instead.)

[taylordowns2000]

This is awesome!!! Is there another issue to modify the text on the search bar and update the docs? I think we agreed that we'd need to:

1. Explain in the documentation why search is currently not supported for very large dataclips.
2. Add an info hover icon to the search bar that explains to the user that they won't be able to find the dataclip they're looking for if it's bigger than ~Xmb. (I know it's hard to estimate, given the difference in index size. But we should try to be as helpful as possible.)