IBM MQ Monitoring Solutions: DevOps Evaluation Guide

This guide evaluates IBM MQ monitoring capabilities across major monitoring platforms from a DevOps perspective. The analysis focuses on operational requirements, technology choices, and practical deployment considerations.

Monitoring Solutions Evaluated

Netdata - Real-time performance monitoring
Datadog - Cloud-based infrastructure monitoring
Dynatrace - Application performance management
Splunk - Data platform with monitoring capabilities
Grafana - Visualization platform (using Prometheus exporter)
Zabbix - Open source enterprise monitoring
Nagios - Infrastructure monitoring and alerting system
CheckMK - Infrastructure monitoring and alerting platform

Solution Abbreviations

For table readability, the following abbreviations are used:

ND = Netdata
DD = Datadog
DT = Dynatrace
SP = Splunk
GR = Grafana
ZB = Zabbix
NG = Nagios
CM = CheckMK

Queue Manager Level Monitoring

Basic Operational Status

Cardinality: 1 metric set per queue manager Technology: PCF MQCMD_INQUIRE_Q_MGR or process monitoring Authority Required: +connect +inq on queue manager Configuration: None Value: Fundamental availability monitoring

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Status	✅	❌	✅	✅	✅	✅	✅	✅	MQIA_Q_MGR_STATUS
Connection count	✅	❌	✅	✅	✅	✅	✅	❌	MQIA_CONNECTION_COUNT
Command level	❌	❌	✅	❌	❌	❌	❌	❌	MQIA_COMMAND_LEVEL
Platform	✅	❌	✅	❌	❌	❌	❌	❌	MQIA_PLATFORM
Uptime	✅	❌	✅	❌	✅	❌	❌	❌	Calculated
Active channels	❌	✅¹	✅	✅	❌	❌	✅	❌	From channel count
Active listeners	❌	❌	❌	❌	✅	❌	❌	❌	MQIA_ACTIVE_LISTENERS
Publish to sub msgs	❌	❌	❌	❌	✅	❌	❌	❌	$SYS topics
Expired msg count	❌	❌	❌	❌	✅	❌	❌	❌	$SYS topics

¹Via channel.channels metric

Resource Utilization

Cardinality: 1 metric set per queue manager Technology: $SYS topic subscriptions (Grafana/Prometheus) or REST API Authority Required: +sub on $SYS topics or REST API access Configuration: Monitor topics must be enabled (MONINT > 0) Value: Capacity planning, performance troubleshooting

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	Collection Method
CPU usage	✅*	❌	❌	❌	✅	❌	❌	❌	$SYS topics
Memory usage	✅*	❌	❌	❌	✅	❌	❌	❌	$SYS topics
Log write latency	❌	❌	❌	❌	✅	❌	❌	❌	$SYS topics
File system usage	❌	❌	❌	❌	✅	❌	❌	❌	$SYS topics
Log bytes used	✅*	❌	✅	❌	✅	❌	❌	❌	See log utilization
Transaction counts	❌	❌	❌	❌	✅	❌	❌	❌	$SYS topics

*Framework implemented, requires MQ 9.0+ with MONINT configured and attribute ID mapping

Queue Manager Log Utilization

Cardinality: 1 metric set per queue manager Technology: PCF commands Authority Required: +inq on queue manager Configuration: None Value: Log space monitoring and capacity planning

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	Description
Log utilization %	❌	❌	✅	❌	❌	❌	❌	❌	6 metrics for log usage

Advanced Queue Manager Attributes

Cardinality: 1 metric set per queue manager Technology: PCF MQCMD_INQUIRE_Q_MGR Authority Required: +connect +inq on queue manager Configuration: None Value: Configuration tracking and compliance

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Distribution lists	❌	✅	❌	❌	❌	❌	❌	❌	MQIA_DIST_LISTS
Max message length	❌	✅	✅	❌	❌	❌	❌	❌	MQIA_MAX_MSG_LENGTH
Max channels	❌	❌	✅	✅	❌	❌	❌	❌	MQIA_MAX_CHANNELS
Active channels	❌	❌	✅	✅	❌	❌	❌	❌	MQIA_ACTIVE_CHANNELS

Queue Level Monitoring

Queue Configuration Attributes

Cardinality: 1 metric set per queue (potentially thousands) Technology: PCF MQCMD_INQUIRE_Q Authority Required: +inq on queues Configuration: None Value: Configuration compliance, capacity planning

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Max depth	✅	✅	✅	✅	✅	✅	❌	❌	MQIA_MAX_Q_DEPTH
Inhibit get/put	✅	✅	✅	❌	❌	❌	❌	❌	MQIA_INHIBIT_GET/PUT
Backout threshold	✅	✅	❌	❌	❌	❌	❌	❌	MQIA_BACKOUT_THRESHOLD
Trigger settings	✅	✅	❌	❌	❌	❌	❌	❌	MQIA_TRIGGER_*
Service interval	✅	✅	❌	✅	❌	❌	❌	❌	MQIA_Q_SERVICE_INTERVAL
Input open option	❌	✅	❌	❌	❌	❌	❌	❌	MQIA_DEF_INPUT_OPEN_OPTION
Depth event config	❌	✅	❌	❌	❌	❌	❌	❌	MQIA_Q_DEPTH_*_EVENT

Queue Runtime Status

Cardinality: 1 metric set per queue Technology: PCF MQCMD_INQUIRE_Q_STATUS Authority Required: +inq on queues Configuration: None Value: Real-time operational monitoring

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Current depth	✅	✅	✅	✅	✅	✅	✅	✅	MQIA_CURRENT_Q_DEPTH
Depth percentage	✅	✅	❌	❌	❌	❌	❌	❌	Calculated
Open input count	✅	✅	✅	✅	✅	✅	❌	✅	MQIA_OPEN_INPUT_COUNT
Open output count	✅	✅	✅	✅	✅	✅	❌	✅	MQIA_OPEN_OUTPUT_COUNT
Oldest message age	✅	✅	✅	✅	✅	✅	✅	✅	MQIACF_OLDEST_MSG_AGE
Queue file size	✅	❌	❌	✅	❌	❌	❌	❌	MQIACF_CUR_Q_FILE_SIZE²
Uncommitted msgs	✅	✅	✅	✅	✅	✅	❌	❌	MQIACF_UNCOMMITTED_MSGS
Last GET time	✅	✅	❌	❌	✅	✅	❌	✅	Time since last GET
Last PUT time	✅	✅	❌	❌	✅	✅	❌	✅	Time since last PUT

²Requires MQ 9.1.5+

Queue Activity Rates (Destructive Operation)

Cardinality: 1 metric set per queue Technology: PCF MQCMD_RESET_Q_STATS Authority Required: +chg on queues Configuration: MONQ(MEDIUM) or MONQ(HIGH) Value: Throughput monitoring Warning: Resets counters after reading

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Enqueue rate	✅³	✅³	❌	✅	✅	❌	❌	❌	MQIA_MSG_ENQ_COUNT
Dequeue rate	✅³	✅³	❌	✅	✅	❌	❌	❌	MQIA_MSG_DEQ_COUNT
High depth	✅³	✅³	❌	✅	❌	❌	❌	❌	MQIA_HIGH_Q_DEPTH

³Disabled by default due to destructive nature

Queue MQI Statistics (Non-Intrusive)

Cardinality: 1 metric set per queue Technology: SYSTEM.ADMIN.STATISTICS.QUEUE subscription Authority Required: +get on SYSTEM.ADMIN.STATISTICS.QUEUE Configuration: STATQ(ON) on queue manager Value: Detailed application behavior analysis

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	Description
Get operations	✅	✅⁴	❌	❌	✅	❌	❌	❌	MQGET count/bytes
Put operations	✅	✅⁴	❌	❌	✅	❌	❌	❌	MQPUT count/bytes
Put1 operations	✅	✅⁴	❌	❌	❌	❌	❌	❌	MQPUT1 count
Get/Put failures	✅	✅⁴	❌	❌	❌	❌	❌	❌	Failed operations
Browse operations	✅	✅⁴	❌	❌	❌	❌	❌	❌	Browse count/bytes
Open/Close ops	✅	❌	❌	❌	❌	❌	❌	❌	MQOPEN/MQCLOSE
Expired messages	✅	✅⁴	❌	❌	✅	❌	❌	❌	Expiration count
Purged messages	✅	✅⁴	❌	❌	❌	❌	❌	❌	Purge count
Non-queued msgs	✅	✅⁴	❌	❌	❌	❌	❌	❌	Direct transfers
Min/Max depth	✅	✅⁴	❌	❌	❌	❌	❌	❌	Queue depth range
Avg queue time	✅	✅⁴	❌	❌	❌	❌	❌	✅	Message latency

⁴Optional feature, disabled by default (collect_statistics_metrics)

Dead Letter Queue Handling

Cardinality: Backout attributes per queue Technology: MQCMD_INQUIRE_Q (configuration attributes) Authority Required: +inq on queues Configuration: Backout requeue name must be configured Value: Message poison prevention and failure handling Note: No solution provides special DLQ monitoring beyond standard queue metrics

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Backout threshold	✅	✅	❌	❌	❌	❌	❌	❌	MQIA_BACKOUT_THRESHOLD
Harden get backout	✅	✅	❌	❌	❌	❌	❌	❌	MQIA_HARDEN_GET_BACKOUT
MQDLH parsing	❌	❌	❌	❌	❌	❌	❌	❌	Reason code extraction
DLQ auto-detection	❌	❌	❌	❌	❌	❌	❌	❌	Special queue handling
Source queue tracking	❌	❌	❌	❌	❌	❌	❌	❌	From MQDLH header

Channel Level Monitoring

Channel Configuration

Cardinality: 1 metric set per channel (tens to hundreds) Technology: PCF MQCMD_INQUIRE_CHANNEL Authority Required: +inq on channels Configuration: None Value: Configuration compliance and tuning

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Batch size	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_BATCH_SIZE
Batch interval	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_BATCH_INTERVAL
Heartbeat interval	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_HB_INTERVAL
Max message length	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_MAX_MSG_LENGTH
Retry settings	✅	✅	✅	❌	❌	❌	❌	❌	MQIACH_*_RETRY
NPM speed	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_NPM_SPEED
Sharing conversations	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_SHARING_CONVERSATIONS

Channel Runtime Status

Cardinality: 1 metric set per running channel Technology: PCF MQCMD_INQUIRE_CHANNEL_STATUS Authority Required: +inq on channels Configuration: None for basic; MONCHL(MEDIUM/HIGH) for detailed Value: Connection health and performance monitoring

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Status	✅	✅	✅	✅	✅	✅	✅	✅	MQIACH_CHANNEL_STATUS
Status summary	❌	❌	✅	❌	❌	❌	❌	❌	Aggregated by status
Messages	✅	✅	✅	✅	✅	❌	❌	❌	MQIACH_MSGS
Bytes sent/rcvd	✅	✅	✅	✅	✅	❌	❌	❌	MQIACH_BYTES_*
Batches	✅	✅	❌	❌	✅	❌	❌	❌	MQIACH_BATCHES
Current convs	✅	❌	❌	✅	❌	❌	❌	❌	Sharing conversations
SSL key resets	✅	✅	❌	❌	❌	❌	❌	❌	MQIACH_SSL_KEY_RESETS
Connection status	❌	✅	❌	❌	❌	❌	❌	❌	MQIACH_CONNS
Active connections	✅	✅	❌	❌	❌	❌	❌	❌	Channel instances
Channel instances	❌	❌	✅	❌	❌	❌	❌	❌	3 instance metrics
In-doubt status	✅	✅	✅	❌	❌	❌	❌	❌	MQIACH_INDOUBT_STATUS
XMITQ time short	✅	❌	✅	❌	✅	❌	❌	❌	MQIACH_XMITQ_TIME_SHORT
XMITQ time long	✅	❌	✅	❌	✅	❌	❌	❌	MQIACH_XMITQ_TIME_LONG
Network time	❌	❌	✅	❌	❌	❌	❌	❌	Channel timing metrics
Exit time	❌	❌	✅	❌	❌	❌	❌	❌	Channel timing metrics
Total time	❌	❌	✅	❌	❌	❌	❌	❌	Channel timing metrics

Channel Statistics (Non-Intrusive)

Cardinality: 1 metric set per channel Technology: SYSTEM.ADMIN.STATISTICS.QUEUE subscription Authority Required: +get on SYSTEM.ADMIN.STATISTICS.QUEUE Configuration: STATQ(ON) and STATCHL(MEDIUM/HIGH) Value: Detailed performance analysis

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	Description
Messages/Bytes	✅	✅⁴	❌	❌	✅	❌	❌	❌	Transfer metrics
Put retries	✅	✅⁴	❌	❌	❌	❌	❌	❌	Retry counts
Batch metrics	✅	✅⁴	❌	❌	❌	❌	❌	❌	Full/partial batches

⁴Optional feature, disabled by default

Topic and Subscription Monitoring

Topic Status

Cardinality: 1 metric set per topic Technology: PCF MQCMD_INQUIRE_TOPIC_STATUS Authority Required: +inq on topics Configuration: Pub/Sub enabled Value: Pub/Sub health monitoring

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Publisher count	✅	⚠️	❌	❌	✅	❌	❌	❌	MQIA_PUB_COUNT
Subscriber count	✅	⚠️	❌	❌	✅	❌	❌	❌	MQIA_SUB_COUNT
Published msgs	✅	⚠️	❌	❌	✅	❌	❌	❌	Via status

⚠️ Documentation claims support but no implementation found

Subscription Status

Cardinality: 1 metric set per subscription (can be very high) Technology: PCF MQCMD_INQUIRE_SUB_STATUS Authority Required: +inq on subscriptions Configuration: Pub/Sub enabled Value: Subscription health and message backlog monitoring

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	PCF Constant
Message count	✅	⚠️	❌	❌	✅	❌	❌	❌	MQIACF_MESSAGE_COUNT
Last message time	✅	⚠️	❌	❌	✅	❌	❌	❌	MQCACF_LAST_MSG_TIME

⚠️ Documentation claims support but no implementation found

Event Monitoring

Event Queue Subscriptions

Cardinality: Events as they occur Technology: Subscribe to SYSTEM.ADMIN.*.EVENT queues Authority Required: +get on event queues Configuration: Enable specific event types on queue manager Value: Real-time alerting for security and operational events

Event Type	ND	DD	DT	SP	GR	ZB	NG	CM	Event Queue
Authority failures	❌	❌	❌	✅	❌	❌	❌	❌	QMGR.EVENT
Queue depth events	❌	❌	❌	✅	❌	❌	❌	❌	QMGR.EVENT
Channel events	❌	❌	❌	✅	❌	❌	❌	❌	CHANNEL.EVENT
Performance events	❌	❌	❌	✅	❌	❌	❌	❌	PERFM.EVENT

Listener Monitoring

Listener Status

Cardinality: 1 metric set per listener (typically 1-5) Technology: PCF MQCMD_INQUIRE_LISTENER_STATUS Authority Required: +inq on listeners Configuration: None Value: Network endpoint availability

Metric	ND	DD	DT	SP	GR	ZB	NG	CM
Status	✅	❌	✅	✅	✅	❌	❌	❌
Port/IP	✅	❌	✅	✅	✅	❌	❌	❌
Backlog	✅	❌	❌	❌	❌	❌	❌	❌

z/OS Specific Monitoring

z/OS Usage Metrics

Cardinality: 1 metric set per queue manager Technology: z/OS specific PCF commands Authority Required: +inq on queue manager Configuration: z/OS platform only Value: z/OS resource usage and performance

Metric	ND	DD	DT	SP	GR	ZB	NG	CM	Description
z/OS CPU usage	❌	❌	✅	❌	❌	❌	❌	❌	8 z/OS metrics
z/OS memory usage	❌	❌	✅	❌	❌	❌	❌	❌	z/OS specific
z/OS paging	❌	❌	✅	❌	❌	❌	❌	❌	z/OS specific

Cluster Monitoring

Cluster Queue Manager Status

Cardinality: 1 metric set per cluster queue manager Technology: PCF MQCMD_INQUIRE_CLUSTER_Q_MGR Authority Required: +inq on cluster Configuration: Cluster must be configured Value: Cluster health and topology monitoring

Metric	ND	DD	DT	SP	GR	ZB	NG	CM
Cluster suspend state	❌	❌	❌	❌	✅	❌	❌	❌
Cluster QM status	❌	❌	❌	❌	✅	❌	❌	❌
Cluster QM availability	❌	❌	❌	❌	✅	❌	❌	❌

Auto-Discovery and Configuration Management

MQ Server Discovery and Instance Management

Technology: Various (network discovery, configuration files, API calls)
Value: Automated setup and dynamic environment adaptation

Capability	ND	DD	DT	SP	GR	ZB	NG	CM	Description
Auto-discover MQ servers	✅	❌	✅	❌	❌	❌	❌	❌	Network discovery of MQ installations
Monitor all auto-discovered Queue Managers	❌	✅	✅	✅	✅	✅	✅	✅	Monitor all QMs found via discovery
Monitor specific Queue Managers	✅	✅	✅	✅	✅	✅	✅	✅	Target specific QM instances
Monitor all auto-discovered Queues	✅	✅⁶	❌	✅	✅	✅⁷	❌	✅	Auto-discover and monitor all queues
Monitor specific Queues	✅	✅	✅	✅	✅	✅	✅	✅	Target specific queue instances
Monitor all auto-discovered Channels	✅	✅⁶	❌	✅	✅	✅⁷	❌	✅	Auto-discover and monitor all channels
Monitor specific Channels	✅	✅	✅	✅	✅	✅	✅	✅	Target specific channel instances
Monitor all auto-discovered Listeners	✅	❌	❌	❌	✅	❌	❌	❌	Auto-discover and monitor all listeners
Monitor specific Listeners	✅	❌	✅	✅	✅	❌	❌	❌	Target specific listener instances
Monitor all auto-discovered Topics	✅	❌	❌	❌	✅	❌	❌	❌	Auto-discover and monitor all topics
Monitor specific Topics	✅	⚠️	❌	❌	✅	❌	❌	❌	Target specific topic instances
Monitor all auto-discovered Subscriptions	✅	❌	❌	❌	✅	❌	❌	❌	Auto-discover and monitor all subscriptions
Monitor specific Subscriptions	✅	⚠️	❌	❌	✅	❌	❌	❌	Target specific subscription instances

⁶Datadog: auto_discover_queues with queue_patterns/queue_regex filtering
⁷Zabbix: Custom auto-discovery templates (e.g., spectroman/ibm-mq-monitoring)

Configuration Management Approaches

Netdata: Network discovery of MQ servers, but requires explicit configuration per queue manager. Supports regex-based filtering for all object types within configured queue managers.

Datadog: Manual server configuration, automatic queue/channel discovery with auto_discover_queues, queue_patterns, and queue_regex for performance optimization. No topic/subscription auto-discovery.

Dynatrace: Automatic IBM MQ process detection via OneAgent, bulk inquiry options, regex-based queue manager filtering. Manual configuration for queue/channel specifics.

Splunk: Manual configuration with IBM WebSphere MQ Modular Input. Supports multiple queues/channels per data input with regex filtering.

Grafana/Prometheus: Manual server configuration, regex-based filtering for <queues>, <channels>, <topics>, <subscriptions> in mq_prometheus exporter configuration.

Zabbix: Manual setup with third-party auto-discovery templates (spectroman/ibm-mq-monitoring) that dynamically create templates for discovered queue managers and objects.

Nagios: Manual configuration per check, supports multiple queues per check with configuration files. Advanced plugins support regex filtering for queue discovery.

CheckMK: Uses dspmq command output for queue manager discovery. Manual configuration for queue-specific monitoring. Limited auto-discovery capabilities.

Collection Characteristics

Update Frequency

Solution	Minimum	Typical	Maximum	Architecture
ND	1s	1s	60s	Edge (per-node)
DD	15s	15-30s	300s	Centralized agent
DT	60s	60s	300s	Centralized agent
SP	10s	60s	300s	Centralized
GR	15s	30-60s	300s	Pull-based
ZB	30s	60-300s	3600s	Centralized
NG	60s	300s	3600s	Centralized
CM	60s	60-300s	3600s	Centralized

Platform Support

Solution	Collector Platform	Remote Support	Architecture
ND	Linux, Windows, macOS	✅ Client mode	CGO-based
DD	Linux, Windows, macOS	✅ Client mode	Python pymqi
DT	Cross-platform	✅ OneAgent	Proprietary
SP	Cross-platform	✅ Client mode	Java-based
GR	Linux, Windows, AIX	✅ Client mode	Go client
ZB	Linux only	❌ Local only	Shell scripts
NG	Linux, Unix, Windows	✅ Client mode	Shell/Perl scripts
CM	Linux, Solaris, AIX, HPUX	✅ Client mode	password-less root ssh

Key Technical Differentiators

Netdata: Highest resolution (1s), complete MQI statistics implementation, Linux-only collector

Datadog: Python-based flexibility, endianness handling for AIX/IBM i, significant gap between documentation and implementation

Dynatrace: Commercial APM integration, z/OS specific metrics, channel timing analysis, topology mapping for transaction tracing

Splunk: Event queue monitoring capability, Java-based cross-platform support

Grafana/Prometheus: $SYS topic subscriptions for resource metrics, multiple export formats, comprehensive z/OS support, cluster monitoring

Zabbix: Local monitoring only, requires sudo access, minimal metric coverage

Nagios: Basic health checks focused on availability monitoring, simple alerting model with command-based checks. Supports queue manager status (qmmon), queue depth (qdmon), channel status (chlmon), message age (mamon), port monitoring (portmon/portcon), process monitoring (brkmon/cmdmon), and log monitoring (fdcmon/refmon) functions

CheckMK: Command-line tool based monitoring using native MQ utilities (dspmq, runmqsc). Supports queue manager availability monitoring, basic queue metrics (depth, message age, last GET/PUT times, open counts, queue time), and channel status monitoring with CRITICAL alerts for STOPPED channels. Limited to metrics available through standard MQ command-line tools.

Service Health Checks

Technology: Various (connection tests, PCF queries, process checks) Value: Proactive alerting and availability monitoring

Check Type	ND	DD	DT	SP	GR	ZB	NG	CM	Description
Connection check	✅	✅	✅	✅	✅	✅	✅	✅	Can connect to QM
Queue manager check	✅	✅	✅	✅	✅	✅	✅	✅	QM responding
Queue availability	✅	✅	✅	✅	✅	✅	✅	✅	Queue accessible
Channel health	✅	✅	✅	✅	✅	✅	✅	✅	Channel status check
Channel status alerts	❌	✅⁵	✅	✅	✅	❌	✅	✅	Status-based alerts

⁵Datadog provides specific service checks with WARNING/CRITICAL based on channel state

Operational Considerations

Authority Requirements: All solutions require similar MQ authorities. Event monitoring requires additional permissions. Datadog requires +chg for reset statistics.
Configuration Impact: Statistics collection (STATQ) has minimal performance impact. Reset operations are destructive.
Cardinality Management: Queue and subscription metrics can create high cardinality. Plan retention accordingly.
Platform Limitations: Only Dynatrace, Splunk, and Grafana offer true cross-platform collectors.
Resolution vs Volume: Higher resolution (Netdata) provides better anomaly detection but increases storage requirements.
Endianness: Datadog provides convert_endianness option for AIX/IBM i platforms.
$SYS Topics: Grafana/Prometheus requires MONINT > 0 on queue manager to enable $SYS topic publications for resource metrics.