Manage agents¶
Users with the Manager role can administer agents as follows:
Agent properties¶
You can configure the agent in order to calibrate its behavior by setting the properties in the agent.config file (per agent).
The configuration file is located at the following path:
<install_dir>/agent/agent.config
You can set these properties from the:
Agent.config file¶
To manually set the properties, navigate to the config file on the relevant server:
<install_dir>/agent/agent.config
The following table describes the properties that you can customize:
| Parameter name | Explanation |
|---|---|
max_dynamic_log_rate | Number of log prints per second |
max_dynamic_log_bytes_rate | Number of bytes per second (on |
log_stats_time_micros | How often to log debugger |
max_condition_cost | Maximum cost in percentage of CPU |
max_log_cpu_cost | Maximum cost of dynamic logging |
max_snapshot_buffer_size | The total size in bytes we allow |
max_snapshot_frame_count | Max frame count we allow to |
breakpoint_expiration_sec | Breakpoint \ Dynamic log |
dynamic_log_quota_recovery_ms | Time to pause dynamic log after |
no_check_certificate | Disable certificate pinning when |
ignore_quota | Disable all performance safe |
exceptions_monitoring_enabled | Enable the capturing of the |
exceptions_monitoring_stdout | Print all of the captured |
exceptions_report_percentage | Process and report only this percentage of the exceptions thrown by the debuggee (A float between 0 and 1.0). |
exceptions_should_report_caught | Report exceptions that were caught by the application as well as exceptions that remained uncaught |
exceptions_max_buffer_size | The total size in bytes we allow to evaluate when capturing an exception (Same as snapshot) |
exceptions_stack_trace_frame_count | Max frame count we allow to collect data from when capturing an exception (Same as snapshot) |
exit_after_report_all | Agent empties queues before the application exits. |
enable_pii_redaction | Enable personally identifiable information (PII) redaction in the agent's side (may have an effect on the application's performance) |
Command line arguments¶
Agent configurations can be changed with a command line argument.
The command line arguments should be inserted after the path of the agent and they should be separated by a comma.
Synopsis¶
-agentpath:<path-to-agent>/lightrun_agent.so=--<parameter>=<value>,--<parameter>=<value>...
Agent parameters¶
Following are the parameters that can be configured from the terminal.
| Parameter name | Explanation | Type |
|---|---|---|
hub_retry_delay_ms | amount of time in milliseconds to sleep before retrying failed requests to Management server | int32 |
debuggee_disabled_delay_ms | amount of time in milliseconds to sleep before checking whether the debugger was enabled back | int32 |
lightrun_extra_class_path | additional directories and files containing resolvable binaries | string |
lightrun_init_wait_time_ms | timeout for wait iflightrun_wait_for_init is set | int32 |
lightrun_wait_for_init | ck application until first time of fetching breakpoints from the server | bool |
enable_safe_caller | allows any method without side effects in expressions | bool |
extra_blocked_methods | additional methods to block for testing purposes | string |
extra_allowed_methods | additional methods allowed for testing purposes | string |
extra_whitelisted_classes | internal names of additional classes to allow for testing purposes | string |
expression_max_classes_load_quota | maximum number of classes that the Nano Java interpreter is allowed to load while evaluating a single breakpoint expression | int32 |
expression_max_interpreter_instructions_quota | maximum number of instructions that the Nano Java interpreter is allowed to execute while evaluating a single breakpoint expression | int32 |
pretty_printers_max_classes_load_quota | maximum number of classes that the Nano Java interpreter is allowed to load while formatting some well known data structures | int32 |
pretty_printers_max_interpreter_instructions_quota | maximum number of instructions that the Nano Java interpreter is allowed to execute while formatting some well-known data structures | int32 |
dynamic_log_max_classes_load_quota | maximum number of classes that the Nano Java interpreter is allowed to load while evaluating all expressions in a single dynamic log statement | int32 |
dynamic_log_max_interpreter_instructions_quota | maximum number of instructions that the Nano Java interpreter is allowed to execute while evaluating all expressions in a single dynamic log | int32 |
safe_caller_max_array_elements | maximum allowed size of the array to copy or allocate in safe caller (copying or allocating larger arrays is considered to be too expensive MISSING | int32 |
safe_caller_max_interpreter_stack_depth | maximum stack depth that safe caller will allow | int32 |
cdbg_description_suffix | additional text to be appended to debug description | string |
cdbg_class_files_cache_size | cache size for class files used in safe method caller | int32 |
cdbg_max_instructions_high | use this value when ignoring quota | int32 |
cdbg_max_stack_depth | maximum number of stack frames to unwind | int32 |
Example
-agentpath:<pathtoagent>/lightrun_agent.so=--enable_safe_caller,--lightrun_init_wait_time_ms=999
View active agents¶
Whenever an agent is running, you can view its details directly from the browser.
Log in to your Lightrun account and navigate to Entities->Agents.
The Agents screen loads, and appears similar to the following:
The following table describes the data:
| Column | Description |
|---|---|
| Id | The unique agent ID |
| Host | The host name of the server on which the agent is running |
| ApiVersion | The Lightrun API version currently implemented |
| Pid | The unique process identifier |
| StartTime | The time at which the agent was started |
| UpdateTime | The last time the agent fetched actions from the server |
| PipeLogs | The current configuration for log piping |
| Labels | Tags attached to this agent |
Last update: April 7, 2021