- Reorganizing installation config settings to move more configurations into UI from environment variables
- Changes to installation config to support premium plans in the enterprise edition
- Fixes the broken premium indicator in account/show and accounts/edit page
The Inboxes::FetchImapEmailsJob is designed to fetch the entire email object and check if its message id is already in the database. However, this process is resource-intensive and time-consuming, as fetching the full email object takes a significant amount of time.
On average, fetching 100 emails can take approximately 3-4 minutes to complete depending on the IMAP server. Since we are not using server flags to identify which emails have already been fetched (to avoid compatibility issues with flags in different email services), we have to fetch all previously available emails. Currently we fetch all the messages that were created from yesterday. This becomes problematic with services like Gmail, which throttle requests based on bandwidth usage.
To address this issue, I have updated the logic as follows:
Fetch the sequence IDs of all the mails in the "Inbox" that were created from yesterday.
Use the FETCH command to fetch only the message-ids using BODY.PEEK[HEADER.FIELDS (MESSAGE-ID)] with the sequence IDs we got in the previous step. This is a faster operation with lower bandwidth usage, as it only returns the sequence ID and message ID.
Check if the message IDs are already present in the database for the selected inbox.
If not present, fetch the entire email object using the FETCH command and create the message.
If the message ID is already present, ignore it and move on to the next message-id.
I have also addressed the issue of duplicate emails appearing in conversations when two fetch email jobs occur simultaneously. I have added a lock for the channel to ensure that the job gracefully exits without waiting for the current job to complete.
Fixes#7247Fixes#6082Fixes#8314
Co-authored-by: Sojan <sojan@pepalo.com>
Introduce the REDIS_SENTINEL_PASSWORD environment variable to customize the behaviour of sentinel passwords.
Co-authored-by: EXT02D22861 <nusret.ozates@consultant.turkcell.com.tr>
Co-authored-by: Sojan Jose <sojan@pepalo.com>
Heroku made some SSL/TLS changes with Redis 6, which is breaking the ActionCable configuration.
Hence providing an environment variable configuration `REDIS_OPENSSL_VERIFY_MODE` to fix that.
set the value `none` for this environment variable in your Heroku installations where breakage occurs.
fixes: #2420
* [#139] Delayed emails for conversations
* Added the setex and get methods to Redis wrapper
* Set the priorities for the sidekiq queues
* Was not able to use mailhog for testing email in local, switched back to letter opener and added comments on using the SMTP settings
* Added after create hood in messages to queue the sending of mail after 2 minutes using sidekiq worker and also set the redis key for the conversation to avoid the email sending for every message
* Added the sidekiq worker to send the email and delete the conversation redis key
* Added the mailer and mail template
* mailer sends the last 10 messages along with the new messages from the time it was queued
* Send email only in development or if smtp config is set
* Send email only in development or if smtp config is set
* Set the SMTP_PORT in production variable
* Adding redis to circle CI
* Specs for the conversation email changes
* Added specs for conversation email sidekiq worker
* Added specs for conversation mailer
* Added specs in message model for the after create hook for notify email
* Send emails only when there is a reply from agent
* set development to use mailhog
* Adding comments for using letter opener
