Juju status gets stuck after large juju run

[Dima Tisnek]

Build the test charm from https://github.com/dimaqq/charm-relation-test
(tag juju-status-stuck-repro @512431c; original code by @pengwyn)

Deploy the charm.

In one terminal, run `watch juju status` or equivalent
In another, run `juju run relation-test/0 set-random-blob size=5000000 buckets=3 repetitions=3 method=ops`

The the action completes, the most recent `juju status` process will be stuck, sleeping seemingly forever.

Simon says:

> that seems like we should have a size limit for things like this
> Mongo has a 16mb document limit and we're blowing past it.
> Hence it's having a hissy fit

Why point fingers at Mongo?
I was occasionally able to get this error in a `juju status` that I ran after the other got stuck:

```
ERROR error trying to stop watcher: hub txn watcher sync error: starting change stream: BSONObj size: 30001016 (0x1C9C778) is invalid. Size must be between 0 and 16793600(16MB) First element: txnNumber: 1
ERROR connection is shut down
```

Essentials:

What the charm does:
```
dima@colima-ahh /c/charm-relation-test (main)> juju run relation-test/0 set-random-blob size=5000000 buckets=3 repetitions=3 method=ops

Running operation 22 with 1 task
  - task 23 on unit-relation-test-0

Waiting for task 23...
09:10:41 Time using ops for size 4.77Mi, bucket key-0, iter 0: 1.946197271347046
09:10:43 Time using ops for size 4.77Mi, bucket key-0, iter 1: 1.9031672477722168
09:10:45 Time using ops for size 4.77Mi, bucket key-0, iter 2: 1.9066450595855713
09:10:47 Time using ops for size 4.77Mi, bucket key-1, iter 0: 1.8808460235595703
09:10:49 Time using ops for size 4.77Mi, bucket key-1, iter 1: 1.8854701519012451
09:10:50 Time using ops for size 4.77Mi, bucket key-1, iter 2: 1.8873159885406494
09:10:52 Time using ops for size 4.77Mi, bucket key-2, iter 0: 1.8849141597747803
09:10:54 Time using ops for size 4.77Mi, bucket key-2, iter 1: 1.9213569164276123
09:10:56 Time using ops for size 4.77Mi, bucket key-2, iter 2: 1.8924317359924316

avg: "1.9009271727667914"
data-gen-time: "0.17358636856079102"
max: "1.946197271347046"
min: "1.8808460235595703"
times: '[1.946197271347046, 1.9031672477722168, 1.9066450595855713, 1.8808460235595703,
  1.8854701519012451, 1.8873159885406494, 1.8849141597747803, 1.9213569164276123,
  1.8924317359924316]'
```

dima@colima-ahh /c/charm-relation-test (main)> juju version
3.5.3-genericlinux-arm64

[Dima Tisnek (dimaqq)]

I think that a similar bug was reported and fixed previously: https://bugs.launchpad.net/juju/+bug/1896209

[Vitaly Antonenko (anvial)]

It looks like we need to rethink the limits (blob size) to sort such cases.

I will mark it as wishlist for now.