Previously we let Shortlink IDs increment endlessly—now we strict it to a limited range and manage ID reservation by keeping track of it.
We also minimized the number of writes and fetches from the chrome API—I was unknowingly hitting hidden API limits that would unpredictably halt execution and make behavior flaky. Now that we deal almost entirely in RAM except for writes and deletes (and a big initial fetch), this has been mitigated.
Previously we let Shortlink IDs increment endlessly—now we strict it to a limited range and manage ID reservation by keeping track of it.
We also minimized the number of writes and fetches from the chrome API—I was unknowingly hitting hidden API limits that would unpredictably halt execution and make behavior flaky. Now that we deal almost entirely in RAM except for writes and deletes (and a big initial fetch), this has been mitigated.