Match string between curly brackets, if including pattern - json

Question

In a log I’d like to match all json objects which type is "sync.out.notify.job.status" and print them. I’m on a Mac (zsh).
I tried: grep -Eo \{.+"sync\.out\.notify\.job\.status".+\}
Mon Jan 11 2021 13:08:46 GMT+0100 (Central European Standard Time) CoreSyncNotifications: sync.out.notify.job.status:
{
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "syncing",
"minorstate": "transferring",
"requestid": "45768456-bd58-4cb5-9dff-dfgjfgdj456",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
Mon Jan 11 2021 13:08:46 GMT+0100 (Central European Standard Time) Vulcan: <- COSY 4.3: sync.out.notify.job.status
Mon Jan 11 2021 13:08:46 GMT+0100 (Central European Standard Time) CoreSyncNotifications: sync.out.notify.job.status:
{
"errors": "",
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "idle",
"minorstate": "complete",
"requestid": "4856658fdgh-efb3-4da9-b5f5-4856658fgj",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
Mon Jan 11 2021 13:08:46 GMT+0100 (Central European Standard Time) WebSocketServer: client_7: SEND syncstate creative_cloud: ok

Answer 1

grep usually searches pattern in single line only unless you use -z option in gnu grep.
You may try this gnu awk command:
awk -v RS='(^|\n){[^}]+?sync.out.notify.job.status[^}]+}(\n|$)' 'RT {printf "%s", RT}' file
{
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "syncing",
"minorstate": "transferring",
"requestid": "45768456-bd58-4cb5-9dff-dfgjfgdj456",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
{
"errors": "",
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "idle",
"minorstate": "complete",
"requestid": "4856658fdgh-efb3-4da9-b5f5-4856658fgj",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}

Answer 2

You can easily and robustly use a JSON-aware tool like jq for this, assuming you want to print the matching blocks:
$ grep -v '^[[:alpha:]]' file | jq 'select(.type=="sync.out.notify.job.status")'
{
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "syncing",
"minorstate": "transferring",
"requestid": "45768456-bd58-4cb5-9dff-dfgjfgdj456",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
{
"errors": "",
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "idle",
"minorstate": "complete",
"requestid": "4856658fdgh-efb3-4da9-b5f5-4856658fgj",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
Here's another way to get from { to } to pipe to jq:
$ sed -n '/^{/,/^}/p' file
{
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "syncing",
"minorstate": "transferring",
"requestid": "45768456-bd58-4cb5-9dff-dfgjfgdj456",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
{
"errors": "",
"cloudid": "00000000-0000-0000-0000-000000000000",
"job": "files",
"majorstate": "idle",
"minorstate": "complete",
"requestid": "4856658fdgh-efb3-4da9-b5f5-4856658fgj",
"type": "sync.out.notify.job.status",
"userid": "XXXXXXXXXXXXXXXXXXXXXXXX#AdobeID",
"version": "1.1.43711"
}
I normally wouldn't suggest using a range expression over using a flag but in this case since it's ALL you're going to do with it, it's probably fine.