Twitter confirms that users are experiencing issues with notifications, likes, and retweets, with some users reporting fluctuating numbers of likes and retweets
Context & Ripple Effects
Twitter's February 2019 confirmation of notification and count glitches reads, against the fuller record, as an early entry in a recurring log: the platform later logged worldwide partial outages hitting login, images, polls, and DMs, a 2021 system irregularity affecting all API v2 endpoints, and a run of 2023 failures including the February 8 outage that left users unable to tweet except by scheduling or via the API.
What sets this incident apart is the symptom: not downtime but drifting engagement numbers — likes and retweets that visibly fluctuate — which strikes at the metrics both users and advertisers treat as signal.
First-order effects
- Users cannot trust the like and retweet counts on their own tweets while the glitch persists, and Twitter's support channels absorb a wave of overlapping reports about notifications.
- Anyone measuring tweet performance during the affected window — social media managers, publishers, journalists tracking virality — gets unstable numbers until the counts settle.
Second-order effects
- Recurring incidents push heavy users toward the workarounds later outages made familiar — scheduled posting and API-based tweeting — shifting routine usage onto paths that bypass the flaky consumer-facing surface.
- Each confirmed episode raises the cost of on-platform engagement counts as a measurement standard, giving third-party analytics and rival platforms an opening to pitch more stable signals.
Third-order effects
- If the cadence holds — partial worldwide outages in 2019, an API-wide irregularity in 2021, multiple global failures in 2023 — reliability itself becomes a competitive variable for a platform whose core product is real-time public conversation.
- Persistent metric instability erodes engagement counts as a shared public currency, pushing advertisers and researchers toward independent measurement layers rather than numbers served by the platform.
The trend: Reliability incidents on Twitter are shifting from rare anomalies to a recurring operational pattern that steadily degrades trust in real-time engagement data.