2025.1.12 (2026-02-25)
Platform:
-
Fix: V8 memory leak
-
Fix: Permission checks
2025.1.11 (2025-06-27)
Platform:
-
Fix: Message history getting deleted when using cb-cli push that excludes message history settings
-
Enhancement: Add prometheus metric
clearblade_rpc_messagesto track total number of RPC messages sent by each node -
Enhancement: Change default
Buckets.QueueEmptyIntervalSecsconfig value to 1 hour
2025.1.10 (2025-06-24)
Platform:
-
Fix: Decode
nullandundefinedbodies in the fetch library instead of throwing an error -
Fix: Remove default throttlers when
EnableDefaultThrottlersis updated -
Fix: Don't allow permissions to be set on invalid MQTT topic filters
-
Fix: verifytables subcommand failing with "table is already a hypertable" and "constraint already exists" errors
-
Fix: system uploads not removing library dependencies from service metadata
-
Fix: Prevent race condition when updating config through the
setsubcommand which could generate malformed toml files -
Fix: Prevent panic when using external brokers in V8.
-
Enhancement: Add prometheus gauge
clearblade_security_sessionsfor user, dev, and device sessions (may cause long scrape times, disabled by default with configSecurity.DisableSessionMetrics) -
Enhancement: Support
{systemKey}wildcard in topic permissions to match the current system
2025.1.9 (2025-06-06)
Platform:
-
Fix: Fix builds for 32-bit edges
2025.1.8 (2025-06-06)
Platform:
-
Enhancement: Allow account registration when using OIDC without requiring full OIDC flow
-
Enhancement: Faster json marshaling on large responses from code services
-
Enhancement: Upgrade Mongo driver to 1.17.4
2025.1.7 (2025-05-27)
Platform:
-
Fix: Setting number values for config values with the
setsubcommand failing on remote cluster nodes due to type conversion error -
Fix: Preloaded request timeout not changed when updated in code service
-
Fix: Fix issue that prevented duktape engines from having their metrics tracked while exiting due to timeout or kill
-
Enhancement: Add
setlocalsubcommand for setting config values on a single cluster node -
Enhancement: Include engine type in currently running services output
-
Enhancement: Add config var
Engine.PreloaderQueueSizeto control preloaded request queue size -
Enhancement: If local queue is full when
MQTT.ShareGroupLocalPreferenceis enabled, forward the message to another node instead of immediately dropping -
Enhancement: Add config var
RPC.RPCConnectionTimeoutto control the initial connection timeout to the platform
2025.1.6 (2025-05-16)
Platform:
-
Fix: Improve V8 engine termination so that services which get stuck in the event loop (E.g. infinite loops) may be killed.
-
Fix: Prevent race condition in AWS Go SDK that could occasionally cause Kinesis MQTT connectors to panic
2025.1.5 (2025-05-12)
Platform:
-
Enhancement: Allow infinite code service restarts, with exponential backoff controlled by engine config variables
QuickRestartThresholdMs,MinRestartIntervalMs, andMaxRestartIntervalMs. Defaults to a 5s threshold, with a minimum interval of 1s and a max of 60s -
Enhancement: Upgrade sqlite dependency to v3.49.1
-
Enhancement: Explicitly close Mongo cursors
-
Enhancement: Add index for email_lower_case column in user tables
-
Enhancement: Reduce the number of logs when a shared queue is full and drops a message
2025.1.4 (2025-04-14)
Platform:
-
Fix: Revise TLS handling to allow TLS on the edge.
-
Enhancement: Add HTTP.UseTLSHTTP config var to enable serving HTTP over TLS
2025.1.3 (2025-04-03)
Platform:
-
Fix: Undo alias resolution to fix relation does not exist error when using an alias the same name as a collection (SQLSTATE 42P01)
2025.1.2 (2025-04-02)
Platform:
-
Fix: Stop substituting aliases in raw queries to avoid table name specified more than once error (SQLSTATE 42712)
2025.1.1 (2025-03-31)
Platform:
-
Fix: Change malloc arg type to support 32-bit architecture builds
2025.1.0 (2025-03-31)
Platform:
-
Fix: Prevent all websocket servers from sharing the same router. The websocket server at the config's
BrokerWSPortserves/mqttandMessagingAuthWSPortserves/mqtt_auth. -
Fix: Allow binary payloads to be sent to code services via webhook request body.
-
Enhancement: Add config option,
DisableCodeCache, which defaults totrue. Due to internal data structure changes, this option must be set totrueduring the blue-green deployment process. -
Enhancement: Add config option,
DisableBrokerCacheto disable redis and use the database to process new MQTT connections and perform client ID checks. Defaults to false -
Enhancement: Change distributed lock wait queue bumper to operate on a per-lock basis instead of per-node
-
Enhancement: Return more detailed error messages when system creation fails.
-
Enhancement: Add prometheus gauge,
clearblade_remote_shell_sessions, to track number of remote shell sessions per edge -
Enhancement: Add prometheus metrics
clearblade_code_services_heap_bytesandclearblade_code_services_heap_bytes_overheadto track long running service heap usage. -
Enhancement: Replace config option
LicenseRenewalWebhookwithLicenseRenewalWebhooksto support fallback URLs. -
Enhancement: Add config variable
ShareGroupLocalPreference(default false) which will distribute messages to share group members on the same node that a message arrives on. -
Enhancement: Add v8 and onnx build support to about response
-
Feature: Support the
clearblade_ailibrary in V8 services -
Feature: Add websocket endpoint to allow opening a remote shell to connected edge devices.
-
Feature: Add RPC message retries for node to node communication. See config options
RPCPlatformMaxRetriesandRPCRetryBackoffSeconds. -
Feature: Track heap usage of javascript engines and report metrics to prometheus. See config option
TrackHeapUsage. -
Feature: Add
cachesubcommand -
Feature: Add the following functions to the
cryptolibrary for working with keys:importKey,generateKey,generateX509Certificate,exportKey. -
Feature: Add edge uptime tracker and metrics reporting