Skip to content
Notifly Notifly

Long-running DB migration progress

The scariest migrations are the ones where it’s unclear for a long time “is it done or another hour?”. An incremental ping every N% or every N minutes helps.

import os, time, requests


def push(t, m, p=4):
    requests.post(f"{os.environ['NOTIFLY_URL']}/message",
                  params={"token": os.environ["NOTIFLY_TOKEN"]},
                  json={"title": t, "message": m, "priority": p}, timeout=5)


total = db.count("legacy_users")
processed = 0
last_push = time.time()
last_pct  = -1


for batch in db.iter_chunks("legacy_users", 10_000):
    migrate(batch)
    processed += len(batch)
    pct = int(processed / total * 100)
    if pct // 10 > last_pct // 10 or time.time() - last_push > 600:
        eta = (time.time() - t0) / max(processed, 1) * (total - processed)
        push(f"🗄️ Migration {pct}%",
             f"{processed:,} / {total:,}\nETA: {int(eta/60)} мин")
        last_push = time.time()
        last_pct  = pct


push("🗄️ Migration: ✅ done", f"Обработано {processed:,} строк за {int((time.time()-t0)/60)} мин", 5)

For migrations performed by an AI agent via apply — wrap them in the agent-run wrapper, and inside call the same push().

Section titled “Related recipes”