telemetry input restructure#4299
Merged
Merged
Conversation
Signed-off-by: Abhishek S A <abhishek.sa3@dell.com>
abhishek-sa1
force-pushed
the
pub/telemetry
branch
from
0abbbbe to
b7c0f37
Compare
priti-parate
approved these changes
Apr 22, 2026
Kratika-P
approved these changes
Apr 23, 2026
abhishek-sa1
added a commit
that referenced
this pull request
May 7, 2026
* vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * adding the cuda and nvidia driver version gate check * victoria log cluster component and VLAgent deployment * updating pod name * storage_config.yml added * Added logic for mounts * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * adding SAN for vicotir agent * powerscale authorization input and input validation for powrscale authorization and powerscale telemetry (#4288) * input validation for powrscale authorization and powerscale telemetry * remove unused messages * fix cide review comments * fix UT issues * input valdiation onlu when powerscale authorization is true * pylint fixes * pylint fixes * update copyright year * Updating SAN changes for vmagengt and vlagent * Updating the SAN for vlagent (#4298) * telemetry input restructure Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> * update telemetry * Update victorialogs-operator-vlcluster.yaml.j2 * Update telemetry.sh.j2 * telemetry input restructure (#4299) * telemetry input restructure Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> * update telemetry * Update victorialogs-operator-vlcluster.yaml.j2 * Update telemetry.sh.j2 --------- Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> * build image fix * build image suffix * Update service_k8s.json Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * issue fix * issue fix * build image fix * Revert "issue fix" This reverts commit 847746c. * Revert changes in telemetry branch (#4308) * Revert "Merge branch 'pub/telemetry' into pub/telemetry" This reverts commit 9b8faf8, reversing changes made to 895354b. * update telemetry fix * Mounts level 1 complete * initial cuda dnf installation * updating arch files * changes with respect to new input structure * powerscale_metrics_enabled variable access * powerscale telemetry enabled metrics access * pylint fixes * changes with respect to new input structure (#4312) * changes with respect to new input structure * powerscale_metrics_enabled variable access * powerscale telemetry enabled metrics access * pylint fixes * vlagent with loadbalancer type * feat(mount_config): enhance per-node bind mounts and secure runcmd execution - Implement literal moustache preservation for cloud-init variables using !unsafe with from_yaml. - Generalize node_key support to allow full metadata paths (e.g., ds.meta_data.instance_data.local_ipv4). - Fix b64encode evaluation in permission and directory creation tasks to ensure valid shell command generation. - Centralize mount-specific runcmd execution in cloud-init templates using base64 encoding to prevent shell injection/parsing issues. - Update storage_config schema to support flexible cloud-init variable references for node_key. - Add safety defaults for mount_on_oim selection to prevent 'attribute not found' errors. - Restore 'union' filter for mounts to maintain idempotency across multiple runs. * remove 9429 as it get added by operator * updating input validation, and default vars for vecotr * remove trailing space * feat(storage,provision): add cloud-init swap support + schema/validation updates New tasks: provision/roles/mount_config/tasks/{swap_config.yml,process_single_swap.yml} Build per-functional-group swap entries in cloud_init_groups_dict Validate swap filename path and size format; include optional maxsize Idempotent merge via combine(recursive=True) Render swap block in cloud-init group templates Inject conditional swap: {filename, size, [maxsize]} across login, compiler, kube (control/node), slurm (control/node) for x86_64/aarch64 Only rendered when cloud_init_groups_dict[fg].swap is present Strengthen input validation (common_validation.py) Enforce no overlap of functional_group_prefix/group across swap entries Validate maxsize >= size (supports bytes or K/M/G/T, and “auto”) Add helper parsers and validators for size comparison Update storage_config schema for swap (storage_config.json) Remove deprecated swap.name field Required: filename, size oneOf: require either functional_group_prefix or group Preserve existing size semantics (supports units/auto) * powerscale syslog support * update copyright year * UT issues * remove TLS mode as it is not supported forsyslog on VLAgent * make syslog source IP as optional * UT issue fixes * VLagent to victorial log transformation * VLAgent to victorial logs cert issue * updating default version * remving vector-idrac * fix for the script does not push the new cert to the Kubernetes secret or restart pods * vast client installation * refactor: make determine_target_groups.yml generic for mounts and swap - Change mount_item reference to config_item parameter - Update process_single_mount.yml to pass config_item: mount_item - Update process_single_swap.yml to pass config_item: swap_item - Enable reuse of target group determination logic This allows both mount and swap processing to use the same task. * single template way * Certificate resolution issue * rsyslog configuration * Added powervault to all * loadbalancer configuration * Added powervault to all functional groups * logs with loadbalncer configuration * get syslog with nodeport * syslog collection work on nodeport * ansible lint fix * update remotewrite url * Verify script * removing audit trail log * login compiler node cuda lock installation * updated the message for nodes where installation skips * Powervaulr iscsi packages added to all nodes * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * revert: temporarily revert discovery fixes for PR workflow * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * slurm cuda,dcg,peermem lock installation * vast enabled flag * fix: remove unsupported service_kube_control_plane_first_x86_64 from SUPPORTED_FUNCTIONAL_GROUPS * adding mod[rpbe for peermem * Victoria defect fixes * victoria changes * fix: Add duplicate mount point validation per functional group - Fix pxe_mapping_file reference in config and validation flow - Add validation to detect duplicate mount points per expanded functional group - Include entry names in error messages for clarity - Update schema and examples for consistency * Revert "revert the vast enabled flag" This reverts commit d20d1fb. * vast conditional installation * refactor: Align storage_config.yml with authoritative schema - Remove extra mount_params profiles (vast_nfs_performance, powervault_iscsi, network_storage, local_storage, bind_mounts, scratch_storage, global) - Keep only essential profiles: default and vast_nfs - Simplify mounts section with two core examples (nfs_slurm, nfs_k8s) - Remove VAST-specific entries and per-node bind mount examples - Comment out powervault_config entries (move to examples) - Comment out swap section (move to examples) - Fix field names: group → groups in comments and examples - Remove vast_enabled flag - Align with storage_config.json schema (only mount_params, mounts, powervault_config, swap sections) - Update comments to reference correct field names (groups instead of group) * comment updated configure-ib-network.sh.j2 Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> * Revereted kube node changes * Removed acidentally committed files * Added ansible.cfg * Ansible lint fixes * lint issue Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> * dns for ib network * adding cuda mount for slurm * adding the dcgm_support * adding variable approach * merge conflict resolution * revert condition * input comment * Adding new files * ansible lint error fixes * feat(discovery): robust OME pagination, GPU collection, and timestamped PXE mapping - Replace @odata.nextLink-based pagination with explicit $top/$skip OData parameters for deterministic, scalable device retrieval (8k-20k+ nodes) - Add configurable page_size parameter (default 200, max 1000) to ome_server_inventory module and wire through Ansible role defaults - Add retry with exponential backoff for transient HTTP 5xx/timeout failures - Return pagination stats (total_devices, page_size, total_pages, pages_fetched) in module output; display summary in playbook - Collect GPU vendor and type from OME devicePciDevice inventory; append GPU_VENDOR and GPU_TYPE columns to PXE mapping CSV - Generate timestamped PXE mapping files (bmc_pxe_mapping_file_<ts>.csv) so each run produces a unique file - Handle OME returning list instead of dict from inventory endpoints - Fix telemetry_config.yml validation: point to telemetry_validation module instead of removed common_validation.validate_telemetry_config - Add 20 unit tests covering pagination, retry, clamping, and filtering * variable chnages wrt new telemetry_config.yml chnages * fix(lint): resolve yaml line-length violations (>160 chars) * feat: implement host-specific mount map in cloud-init templates - Enable host-specific mount map generation from PXE mapping data - Add vendor_data metadata structure to cloud-init runcmd for dynamic mounts - Update cloud-init variable syntax from Jinja2 template to cloud-init query - Fix variable reference in build_host_mount_map.yml (PXE_GROUP -> GROUP_NAME) - Change mkdir commands to verbose mode (mkdir -pv) - Update pxe_mapping_file_path to use omnia_shared_path with path normalization This allows per-host mount configurations to be dynamically applied during node provisioning via cloud-init vendor_data metadata. * Removed group map removed legacy code * Update logical_validation.py- remove repetitive import * Update telemetry_pod_cleanup.yaml.j2 - revert telemetry pod cleanup changes * Update idrac_telemetry_statefulset.yaml.j2 - revert statefulset changes * Update telemetry_prereq.yml-remove unnecessary comment * Update csi_driver_powerscale.json- update URL * Update service_k8s.json-update version as per release 1.16.3 * Update telemetry_config.yml- fix csm observability values.yaml version * Update idrac_telemetry_statefulset.yaml.j2-update variable parsing * Update configure_powerscale_syslog.sh.j2 -fix syntax error for syslog * removing dcgm_support variable * removing stale entries * remove default dns entry for ib network Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> * image layer initial commit * minor chnages * mounting hpc_tools * removinf from control node * Fix mount ordering and bind mount creation for cloud-init runcmd - Restructure runcmd to mount NFS shares before creating bind mount sources - Add explicit mount commands after each NFS fstab entry to ensure mounts are available before subdirectories are created on them - Add mkdir commands for bind mount target directories on mounted NFS - Consolidate mount -av at appropriate checkpoints to avoid premature mounting before all fstab entries and directories are ready - Update process_single_mount.yml to add mkdir for bind mount targets before fstab entries - Add "Refresh mounts" task to append final mount -av after all setup This ensures the correct sequence: 1. Add NFS mount to fstab 2. Mount NFS share 3. Create subdirectories on mounted NFS 4. Add bind mount entries to fstab 5. Mount bind mounts Fixes mount failures: "special device does not exist" and "mount point does not exist" errors during cloud-init execution. * VMscraper to scrape metrics from OTEL collector (#4362) * pub telemetry changes * service to scrape metrics from OTEL collector * vmservice to scrape metrics from otel collector * update endpoints * revert other changes * revert merge changes as per head * revert variable set * revert changes * revert changes * pylint fixes * ansible lint fixes * updating completion messaage * telemetry validation while prepare oim * update condition * added check for LDMS * Fix: Add retry logic for ochami node discovery with SMD readiness checks - Extract service readiness checks into separate block/rescue pattern - Add SMD API health check before discovery attempt - Implement automatic retry on discovery failure with service restart - Increase service check timeout from 2 to 2 minutes (12 retries × 10s delay) - Prevent connection refused errors by ensuring SMD endpoint is ready This addresses the race condition where systemd marks smd service as "started" but the HTTP endpoint at oimcp.oim.test:8443 isn't accepting connections yet. * Update smd api check command * chnaged approach * Cloud_init and other service failure messages * Updated messages * fix: BMC discovery credential prompting and PARENT_SERVICE_TAG assignment (#4366) 1. Skip OME credential prompt when enable_bmc_discovery is false - Changed discovery credentials from mandatory to conditional_mandatory in credential utility vars (gated on enable_bmc_discovery) - Added set_fact in prepare_oim.yml to promote enable_bmc_discovery from namespaced to top-level scope before credential utility runs 2. Fix PARENT_SERVICE_TAG assignment in PXE mapping - Source changed from service_kube_control_plane to service_kube_node - Only slurm_node_aarch64 and slurm_node_x86_64 receive PARENT_SERVICE_TAG - All other roles (control_plane, kube_node, login, slurm_control) remain empty * changing the script a little * ansible lint * ansible lint * Update provision_mapping_nodes.yml Signed-off-by: Jagadeesh N V <39791839+jagadeeshnv@users.noreply.github.com> * Refactor: Update K8s NFS configuration to use unified mounts structure - Replace nfs_client_params with mounts list for consistency - Extract k8s_nfs_server_ip from mounts source field (split on ':') - Extract k8s_client_mount_path from mounts mount_point field - Introduce k8s_nfs_server_path as unified source reference - Update error message to use k8s_nfs_server_path for clarity - Comment out deprecated k8s_server_share_path and nfs_server_ip references - Consolidate 'mount -av' into single set_fact call instead of separate task - Remove redundant debug task for single_mnt_runcmd - Add loop_control label for better Ansible output readability - Improve code formatting for regex_replace operations (no functional change) - Streamline permission and bind mount runcmd list construction * Updated xlurm custom repo names * Updated repo names * Removed static mounts block * rearrange args * Removed unecessary mounts * Merge pull request #4371 from Kratika-P/pub/q2_dev Telemetry enabled flag set for creating nd executing the telemetry.sh * fix(provision): resolve K8s NFS mount race condition and improve Slurm support detection - Move mkdir for K8s NFS mount directories before fstab entries to prevent race condition where mount attempts occur before target directories exist - Fix duplicate mount point names in storage_config.yml (ps3: slurm_login -> slurm_login_kube_node, ps4: login_node -> login_kubectrl) to avoid conflicts between functional groups - Add mkdir -p for /tmp/crio-storage, /var/lib/etcd, /var/lib/kubelet, /etc/kubernetes, /var/log/pods, /var/lib/packages BEFORE adding NFS entries to /etc/fstab - Affected templates: - ci-group-service_kube_control_plane_first_x86_64.yaml.j2 - ci-group-service_kube_control_plane_x86_64.yaml.j2 - ci-group-service_kube_node_x86_64.yaml.j2 - Remove redundant 'mount -av' from bind mount runcmd sequence (process_single_mount.yml) - Functional groups now handle mount -av independently - Fix slurm_support detection logic to check both software_config AND functional_groups - Initialize slurm_support to false, then set true only if: 1. slurm_custom is in software_config.softwares, AND 2. slurm_control_node_x86_64 exists in functional_groups - Apply fix to both mount_config and slurm_config roles - Remove unused slurm_support var from mount_config/vars/main.yml - Remove verbose flags from swap commands (fallocate -v, chmod -v, mkswap -v, swapon -v) in ci-group-login_compiler_node_aarch64.yaml.j2 - Update storage_config.json schema to reflect new mount point naming conventions This commit addresses Issue #3 (race condition) and Issue #4 (missing directory creation) from the cloud-init analysis, preventing silent NFS mount failures on K8s nodes. * Added missing mount-a * papi url chnaged * Updated powervault to service_kube nodes * LDAP home mount updated * Input refinement * Updated descriptions * gpu defect- peermem defect- prolog removal * defect fix for pxe validation * lint fixes * Templated swap part --------- Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> Signed-off-by: sakshi-singla-1735 <sakshi.s@dell.com> Signed-off-by: Jagadeesh N V <39791839+jagadeeshnv@users.noreply.github.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: mcas <sakshi.s@dell.com> Co-authored-by: Jagadeesh N V <jagadeesh_n_v@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit.jadhav@dell.com> Co-authored-by: Jagadeesh N V <39791839+jagadeeshnv@users.noreply.github.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
telemetry input restructure