10 comments.

  1. Dumais

    only difference from normal response is that that empty array response does not have this header Content-Encoding: zstd

  2. Dumais

    tried with debug=all when returning empty data array "debug" field is not set at all ( it is set when request succeeds )

  3. Michele

    We have also noticed the same anomaly: https://developers.facebook.com/community/threads/1363475634337488/

    1. Michele

      We have also noticed the same anomaly: https://developers.facebook.com/community/threads/1363475634337488/

      1. Michele

        We have also noticed the same anomaly: https://developers.facebook.com/community/threads/1363475634337488/

    2. Michele

      We have also noticed the same anomaly: https://developers.facebook.com/community/threads/1363475634337488/

  4. Leon

    Also noticing this. Timeline in PST: ~1:50pm Feb 26th. Started at low volume and remained consistent ~10:10am Feb 27th. Dropped zero (no issues) ~ll:25am Feb 27. Ramped up to high volume, this is when it started getting really bad and our own monitoring detected it. It has been at this level since.
    Currently, if I do 100 identical calls to this endpoint with the same user token I get a failure (empty data array) rate of between 56% and 73% which is quite bad. I've done about 5 tests with 100 calls each to get these results. Over the last hour.

    1. Leon

      Also noticing this. Timeline in PST: ~1:50pm Feb 26th. Started at low volume and remained consistent ~10:10am Feb 27th. Dropped zero (no issues) ~ll:25am Feb 27. Ramped up to high volume, this is when it started getting really bad and our own monitoring detected it. It has been at this level since.
      Currently, if I do 100 identical calls to this endpoint with the same user token I get a failure (empty data array) rate of between 56% and 73% which is quite bad. I've done about 5 tests with 100 calls each to get these results. Over the last hour.

      1. Leon

        Also noticing this. Timeline in PST: ~1:50pm Feb 26th. Started at low volume and remained consistent ~10:10am Feb 27th. Dropped zero (no issues) ~ll:25am Feb 27. Ramped up to high volume, this is when it started getting really bad and our own monitoring detected it. It has been at this level since.
        Currently, if I do 100 identical calls to this endpoint with the same user token I get a failure (empty data array) rate of between 56% and 73% which is quite bad. I've done about 5 tests with 100 calls each to get these results. Over the last hour.

    2. Leon

      Also noticing this. Timeline in PST: ~1:50pm Feb 26th. Started at low volume and remained consistent ~10:10am Feb 27th. Dropped zero (no issues) ~ll:25am Feb 27. Ramped up to high volume, this is when it started getting really bad and our own monitoring detected it. It has been at this level since.
      Currently, if I do 100 identical calls to this endpoint with the same user token I get a failure (empty data array) rate of between 56% and 73% which is quite bad. I've done about 5 tests with 100 calls each to get these results. Over the last hour.

  5. Leon

    Ok, have had confirmation form a Meta contact that it's been fixed. My tests confirm that 100% of /token_debug calls now work.

    1. Leon

      Ok, have had confirmation form a Meta contact that it's been fixed. My tests confirm that 100% of /token_debug calls now work.

Add a new comment.