◢███◤      ◢██◤                            ◢██◤                            
     ◢██◤       ◢██◤                            ◢██◤                             
    ◢██◤       ◢██◤                            ◢██◤                              
   ◢██◤       ◢██◤                            ◢██◤                               
  ◢██◤       ◢██◤                            ◢██◤                                
◢███◤       ◢██◤                            ◢██◤                          ◥██◣   
◥███       ◢█████◣    ◢████████◤ ◢███████◤ ◢██◤ ◢██◤                 ◢██◤   ██◣  
 ███      ◢███████◣        ◢██◤ ◢██◤ ◢██◤ ◢███████◤                 ◢██◤    ███  
 ███     ◢██◤  ◢██◤ ◢████████◤ ◢██◤      ◢█████◣                   ◢██◤     ███  
 ███    ◢██◤  ◢██◤ ◢██◤  ███◤ ◢██◤ ◢██◤ ◢██◤◥███◣                 ◢██◤      ███  
 ◥██   ◢██◤  ◢██◤ ◢████████◤ ◢███████◤ ◢██◤  ◥███◣               ◢██◤       ███◣ 
  ◥██◣                                                          ◢██◤       ◢███◤ 
                                 ◢███◤ ◢███◤ ◢██◤  ◢██◤ ◢█████████◤       ◢██◤   
                                ◢█████████◤ ◢██◤  ◢██◤ ◢██◤  ████◤       ◢██◤    
                               ◢██◤◢█◤◢██◤ ◢██◤  ◢██◤ ◢██◤   ███◤       ◢██◤     
                              ◢██◤   ◢██◤ ◢████████◤ ◢█████████◤       ◢██◤      
                             ◢██◤   ◢██◤ ◢████████◤ ◢█████████◤      ◢███◤       

0007
[Request] context.run_stats
xb2kzo
I'd love to see more statistics on parse/etc times in script runs, for $reasons. What I'd love to see is an object in context with the following properties: { `Nqueue_time`: the time the script run sat in the queue (i.e between when the server first received the request and when a worker first started doing actual work) `Nlookup_time`: the time spent fetching source from the db, but excluding as much assembly as possible `Nparse_time`: the time spent assembling and parsing the script (including esprima) } if just giving us raw dates between steps (i.e. `Nentered_queue`, `Nbegan_lookup`, `Nbegan_parse`, `C_START`) is easier, that's also fine, we can do our own subtraction. We just need some metrics to know WHAT STEP in the process is the problem one so we can fix it. Guessing blindly hasn't worked. This could only be given to the top-level script, or copied into each context arg all the way down. Either works. Alternatively, a trust script that returns equivalent data for a run (`Ftrust`.`Lrun_stats`? `Fscripts`.`Lparse_stats`?) would be fine. We just need the data somehow.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -