September 2024

Get Notified When Devices In Your Fleet Go Offline

Our new Device Offline Alert triggers notifications when a monitored Device hasn't communicated with the Memfault platform for a specified time period. You can configure the alert to send a message via email, to a Slack channel, or configure webhooks to use wherever else you need to be alerted. Device Offline Alerts are assessed every 5 minutes and can be tailored with start and end delays to manage noise and avoid premature alerts due to brief connectivity losses. Read the docs for more details on how to setup and configure Device Offline Alerts in Memfault.

Processing Log History Now Available For 60 Days

We extended our Processing Log retention period from 7 to 60 days. Memfault's Processing Log provides a chronological view of events and data essential for diagnosing SDK integration issues and tracking data uploads. You can filter on date range, Device serial, software version, log type, and level to quickly locate relevant entries. You can also access detailed metadata on a log, and focus on specific transactions, aiding in a streamlined diagnostic investigation. Read the docs for more detail on how to best use the Processing Log for troubleshooting Device behavior and SDK integrations in Memfault.

General

Added:

• Closed beta of Log Management and Analytics for fleet-wide search and analysis of log data, with tabular logs and distribution charts.
• Ability to create alerts for when devices go offline, with configuration options for Incident Start and End Delay.
• The ability to set device attribute values via CSV upload.
• A status filter to the Release Activations page, allowing users to view only active or pulled releases.
• Ability for users to see their email verification status and re-send user email verification links from their profile.
• An emoji picker feature for dashboards.
• A new Organization switcher in the navigation menu.
• Informational warning that slugs are immutable when creating a new cohort.
• Link to the Webhook documentation for easier access through the project notification settings.

Improved

• Y-axis scaling for bar charts to ensure the smallest bar is always visible.
• The timeline search now allows users to locate off-screen metrics using native browser search functionality.
• Updated the Limited OS Version Data warning text to specify the date before which OS Version data is unavailable.
• Increased the historical time range for the Processing Log from 7 days to 60 days.

Changed

• Restricted Trigger Push OTA permissions to Project Managers to align with the access control policy.

Fixed

• An issue where deleted Devices were incorrectly included in the total number of Devices returned in a Device search.
• An issue where the OS Version filter was not copied over to the Issue Search when drilling down from the Trace/Issues chart.

MCU

SDK versions 1.11.3, 1.11.4, 1.11.5, and 1.12.0 were released. Some highlights:

Added:

• Introduced two new core metrics, MemfaultSdkMetric_os_name and MemfaultSdkMetric_os_version, to capture OS/platform name and version.
• Zephyr: Introduced built-in WiFi metrics for connected time, disconnect count, and AP OUI, enabled by default on systems with WiFi.
• ESP-IDF: Added options for uploading logs by default when using MEMFAULT_HTTP_PERIODIC_UPLOAD, and a new metric memory_pct_max to capture max heap usage percentage.

Improved:

• ESP-IDF: Improved CLI output for post_chunks command to be more informative.

Changed:

• ESP-IDF: Renamed cpu_temp metric to thermal_cpu_c for clarity. Adjusted Kconfig settings for heap metrics.

Android

SDK versions 5.0.0 and 5.1.0 were released. Some highlights:

Added:

• New metrics for screen on/off battery drain, including battery_screen_on_discharge_duration_ms, battery_screen_on_soc_pct_drop, battery_screen_off_discharge_duration_ms, and battery_screen_off_soc_pct_drop.
• Storage of software version alongside metric reports to ensure correct association with software updates.

Improved:

• Metrics collection process, with immediate collection after OTA updates to ensure accurate reporting.

Fixed:

• Stability Device Vitals issue where operational_hours was incorrectly set to zero.

Linux

SDK version 1.15.0 was released. Some highlights:

Added:

• A new per-process operational_crashes metric, allowing for detailed crash tracking by process.
• A built-in log level extractor for logs ingested by memfaultd, enabling regex-based log level detection.
• New built-in metrics for CPU, memory, and storage usage by process, as well as network connectivity metrics.

Fixed:

• Fixed a bug where metrics captured before data collection activation were incorrectly uploaded after enabling data collection.

Changed:

• System metric collection is now enabled by default in memfaultd, replacing some collectd metrics with new keys.
• Log messages from memfaultd now consistently include severity levels.

CLI

We did not publish a new version of the memfault-cli in September 2024.