What is a proper retry interval and retry time for Canton related error

Just hope to check if there’s a reasonable or recommended time interval between each retry and max time for retries if canton transaction failed (e.g. sequencer timeout, max rate exceeded etc…)

I understand this should depend heavily on different applications and with different context, just want to see if there’s a number to start with.

Thanks a lot in advance.

Hi Judy,

A reference to retries to the Sequencer is in the Domain Connectivity Service documentation.

Specifically:

message DomainConnectionConfig {
    // participant local identifier of the target domain
    string domain_alias = 1;
    // connection information to sequencer
    com.digitalasset.canton.protocol.v0.SequencerConnection sequencerConnection = 2;
    // if false, then domain needs to be manually connected to (default false)
    bool manual_connect = 3;
    // optional domainId (if TLS isn't to be trusted)
    string domain_id = 4;
    // optional. Influences whether the participant submits to this domain, if several domains are eligible
    int32 priority = 5;
    // initial delay before an attempt to reconnect to the sequencer
    google.protobuf.Duration initialRetryDelay = 6;
    // maximum delay before an attempt to reconnect to the sequencer
    google.protobuf.Duration maxRetryDelay = 7;
    // configuration for how time is tracked and requested on this domain
    com.digitalasset.canton.time.admin.v0.DomainTimeTrackerConfig timeTracker = 8;

As far as a specific period between one retry and another, I cannot see a minimum value but will keep looking.

1 Like

Thanks so much @Ben_M for the swift reply.

I was actually thinking about application (ledger client) retries.
But from the config below, I wonder if it’s recommended to retry from participant, instead from the application.
Also is there a recommended time of retries?
(I only see initialRetryDelay and maxRetryDelay, maybe I missed it)

Thanks so so much again!

If there is a defined or recommended number of retries, I have not seen it yet.

Noted, thanks still @Ben_M!