def graph_num_clients(step_size, output_name, title, data):
"""
Create a stacked graph per client pool.
Arguments:
output_name(Path): where to write the graph
title(str): title for the graph
step_size(int): number of milliseconds between samples
data(DataFrame): data to graph
"""
max_time = data['end'].max()
clients = data['client'].unique()
xs = list()
ydata = dict() # client -> data
time_step = 0
while time_step < max_time:
xs.append(map_utils.timestamp_to_minutes(time_step))
for client in sorted(clients):
ys = ydata.get(client, list())
y_value = data.loc[(data['start'] <= time_step)
& (time_step < data['end']) &
(data['client']
== client)]['num clients'].sum()
ys.append(y_value)
ydata[client] = ys
time_step = time_step + step_size
fig, ax = map_utils.subplots()
ax.set_title(title)
ax.set_xlabel("Time (minutes)")
ax.set_ylabel("Number of clients")
ax.set_xlim(left=0, right=map_utils.timestamp_to_minutes(max_time))
stack_labels, stack_ys = zip(*(sorted(ydata.items())))
ax.stackplot(xs, stack_ys, labels=stack_labels)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles,
labels,
bbox_to_anchor=(1.04, 1),
loc="upper left")
fig.savefig(output_name.as_posix(),
format='png',
bbox_extra_artists=(lgd, ),
bbox_inches='tight')
plt.close(fig)
def output_graph(output, data, min_time):
fig, ax = map_utils.subplots()
ax.set_title("Resource report generation lag")
ax.set_xlabel("Time (minutes)")
ax.set_ylabel("Difference (seconds)")
max_minutes = 0
for node_name, plot_data in data.items():
(xs, ys) = plot_data
minutes = [map_utils.timestamp_to_minutes(x - min_time) for x in xs]
max_minutes = max(max_minutes, max(minutes))
ax.scatter(minutes, ys, label=node_name, s=1)
ax.set_xlim(left=0, right=max_minutes)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles,
labels,
bbox_to_anchor=(1.04, 1),
loc="upper left")
output_name = output / 'resource-report-time-lag.png'
plt.savefig(output_name.as_posix(),
format='png',
bbox_extra_artists=(lgd, ),
bbox_inches='tight')
plt.close(fig)
def relative_timestamps_service(data, min_timestamp):
"""
Arguments:
data (dict): service -> timestamp -> ServiceCounts
min_timestamp (int): first timestamp as milliseconds since the Epoch
Returns:
dict: service -> relative minutes -> ServiceCounts
"""
relative_data = dict()
for service, service_data in data.items():
relative_service_data = dict()
for timestamp, counts in sorted(service_data.items()):
minutes = map_utils.timestamp_to_minutes(timestamp - min_timestamp)
relative_service_data[minutes] = counts
relative_data[service] = relative_service_data
return relative_data
def output_graph(output, ncp, xs, ys, first_timestamp_ms):
"""
Arguments:
output(Path): output directory
ncp(str): name of the NCP the graph is for
xs: list of x-values
ys(dict): neighbor to y-values
first_timestamp_ms(int): initial timestamp to subtract from values
"""
fig, ax = map_utils.subplots()
ax.set_title(f"RLG Resource report lag for node {ncp}")
ax.set_xlabel("Time (minutes)")
ax.set_ylabel("Difference (seconds)")
minutes = [
map_utils.timestamp_to_minutes(x - first_timestamp_ms) for x in xs
]
max_minutes = max(minutes)
for node_name, plot_data in ys.items():
get_logger().debug(
"Graphing NCP %s with neighbor %s. x.len: %d y.len: %d", ncp,
node_name, len(minutes), len(plot_data))
ax.scatter(minutes, plot_data, label=node_name, s=1)
ax.set_xlim(left=0, right=max_minutes)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles,
labels,
bbox_to_anchor=(1.04, 1),
loc="upper left")
output_name = output / f'rlg-node-resource-report-time-lag_{ncp}.png'
plt.savefig(output_name.as_posix(),
format='png',
bbox_extra_artists=(lgd, ),
bbox_inches='tight')
plt.close(fig)
def output_region_capacity_graph_for_app(first_timestamp, output, capacity, app):
"""
Create a graph per node attribute for the specified service with a series for each region.
Creates files with names capacity-<attribute>-<app>.png
Args:
first_timestamp(datetime.datetime): when the simulation started
output (Path): output directory
capacity (boolean): if true output a graph of capacity only
app (str): name of the service to generate the graph for
"""
first_timestamp_ms = first_timestamp.timestamp() * 1000
max_minutes = 0
try:
label='capacity'
all_regions = set(capacity.keys())
all_attrs = set()
for region, region_data in capacity.items():
all_attrs.update(region_data.keys())
for attr in all_attrs:
if 'QueueLength' == attr:
# graphing queue length capacity doesn't work well because we've hardcoded it to be a big number
continue
frames = list()
fig, ax = map_utils.subplots()
ax.set_title(f"{attr} {label} for {app}")
ax.set_xlabel("Time (minutes)")
stack_xs = None
stack_ys = list()
stack_labels = list()
for region in sorted(all_regions):
region_data = capacity.get(region, dict())
attr_data = region_data.get(attr, dict())
app_data = attr_data.get(app, dict())
if len(app_data) > 0:
pairs = sorted(app_data.items())
timestamps, values = zip(*pairs)
times = [map_utils.timestamp_to_minutes(float(t) - first_timestamp_ms) for t in timestamps]
max_minutes = max(max_minutes, max(times))
series_label=f"{region} capacity"
frames.append(pd.DataFrame(list(values), index=list(times), columns=[series_label]))
ax.plot(times, values, label=series_label)
if stack_xs is None:
stack_xs = times
stack_ys.append(values)
stack_labels.append(series_label)
ax.set_xlim(left=0, right=max_minutes)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, bbox_to_anchor=(1.04, 1), loc="upper left")
output_name = output / f"{label}-{attr}-{app}.png"
fig.savefig(output_name.as_posix(), format='png', bbox_extra_artists=(lgd,), bbox_inches='tight')
plt.close(fig)
# write out CSV of the plot data
df = pd.concat(frames, axis=1)
df.to_csv(output / f"{label}-{attr}-{app}.csv", index_label="relative minutes")
# create stacked plot
fig, ax = map_utils.subplots()
ax.set_title(f"{attr} {label} for {app}")
ax.set_xlabel("Time (minutes)")
ax.set_xlim(left=0, right=max_minutes)
ax.stackplot(stack_xs, stack_ys, labels=stack_labels)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, bbox_to_anchor=(1.04, 1), loc="upper left")
output_name = output / f"{label}-{attr}-{app}_stacked.png"
fig.savefig(output_name.as_posix(), format='png', bbox_extra_artists=(lgd,), bbox_inches='tight')
plt.close(fig)
except:
get_logger().exception("Unexpected error")
def output_region_graphs_for_app(first_timestamp, output, capacity, data, label, app):
"""
Create a graph per node attribute for the specified app with a series for each region.
Creates files with names <label>-<attribute>-<app>.png
Args:
first_timestamp(datetime.datetime): when the simulation started
output (Path): output directory
data (dict): region -> attr -> app -> timestamp -> value
label (str): used to label the graph and generate the filenames
capacity (dict): region -> attr -> app -> timestamp -> value
app (str): service to generate the graph for
"""
first_timestamp_ms = first_timestamp.timestamp() * 1000
max_minutes = 0
try:
all_regions = set(capacity.keys())
all_attrs = set()
for region, region_data in data.items():
all_attrs.update(region_data.keys())
for attr in all_attrs:
stack_xs = None
stack_ys = list()
stack_labels = list()
frames = list()
fig, ax = map_utils.subplots()
ax.set_title(f"{attr} {label} for {app}")
ax.set_xlabel("Time (minutes)")
for region in sorted(all_regions):
region_data = data.get(region, dict())
attr_data = region_data.get(attr, dict())
capacity_region = capacity.get(region, dict())
if attr in capacity_region and 'QueueLength' != attr:
app_data = capacity_region[attr].get(app, dict())
if len(app_data) > 0:
pairs = sorted(app_data.items())
timestamps, values = zip(*pairs)
times = [map_utils.timestamp_to_minutes(float(t) - first_timestamp_ms) for t in timestamps]
max_minutes = max(max_minutes, max(times))
series_label=f"{region} capacity"
frames.append(pd.DataFrame(list(values), index=list(times), columns=[series_label]))
ax.plot(times, values, label=series_label)
app_data = attr_data.get(app, dict())
if len(app_data) > 0:
pairs = sorted(app_data.items())
timestamps, values = zip(*pairs)
times = [map_utils.timestamp_to_minutes(float(t) - first_timestamp_ms) for t in timestamps]
max_minutes = max(max_minutes, max(times))
frames.append(pd.DataFrame(list(values), index=list(times), columns=[region]))
ax.plot(times, values, label=region)
if stack_xs is None:
stack_xs = times
stack_ys.append(values)
stack_labels.append(region)
ax.set_xlim(left=0, right=max_minutes)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, bbox_to_anchor=(1.04, 1), loc="upper left")
output_name = output / f"{label}-{attr}-{app}.png"
fig.savefig(output_name.as_posix(), format='png', bbox_extra_artists=(lgd,), bbox_inches='tight')
plt.close(fig)
# write out CSV of the plot data
df = pd.concat(frames, axis=1)
df.to_csv(output / f"{label}-{attr}-{app}.csv", index_label="relative minutes")
# create stacked plot
fig, ax = map_utils.subplots()
ax.set_title(f"{attr} {label} for {app}")
ax.set_xlabel("Time (minutes)")
ax.set_xlim(left=0, right=max_minutes)
ax.stackplot(stack_xs, stack_ys, labels=stack_labels)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, bbox_to_anchor=(1.04, 1), loc="upper left")
output_name = output / f"{label}-{attr}-{app}_stacked.png"
fig.savefig(output_name.as_posix(), format='png', bbox_extra_artists=(lgd,), bbox_inches='tight')
plt.close(fig)
except:
get_logger().exception("Unexpected error")
def output_graphs_per_region(first_timestamp, output, capacity, data, label):
"""
Create a graph per node attribute.
Creates files with names <label>-<attribute>-<region>.png
Args:
first_timestamp(datetime.datetime): when the simulation started
output (Path): output directory
data (dict): region -> attr -> app -> timestamp -> value
label (str): used to label the graph and generate the filenames
capacity: region -> attr -> app -> timestamp -> value
"""
first_timestamp_ms = first_timestamp.timestamp() * 1000
max_minutes = 0
try:
for region, region_data in data.items():
capacity_region = capacity[region]
for attr, attr_data in region_data.items():
frames = list()
fig, ax = map_utils.subplots()
ax.set_title(f"{attr} {label} in {region}")
ax.set_xlabel("Time (minutes)")
stack_xs = None
stack_ys = list()
stack_labels = list()
total_capacity = None
if attr in capacity_region and 'QueueLength' != attr:
for app, app_data in sorted(capacity_region[attr].items()):
if len(app_data) > 0:
pairs = sorted(app_data.items())
timestamps, values = zip(*pairs)
times = [map_utils.timestamp_to_minutes(float(t) - first_timestamp_ms) for t in timestamps]
max_minutes = max(max_minutes, max(times))
series_label=f"{app} capacity"
frames.append(pd.DataFrame(list(values), index=list(times), columns=[series_label]))
ax.plot(times, values, label=series_label)
if total_capacity is None:
total_capacity = app_data.copy()
else:
total_capacity = {k: total_capacity.get(k, 0) + app_data.get(k, 0) for k in set(total_capacity) | set(app_data)}
app_timestamps = None
for app, app_data in sorted(attr_data.items()):
if len(app_data) > 0:
pairs = sorted(app_data.items())
timestamps, values = zip(*pairs)
times = [map_utils.timestamp_to_minutes(float(t) - first_timestamp_ms) for t in timestamps]
max_minutes = max(max_minutes, max(times))
if app_timestamps is None:
app_timestamps = timestamps
frames.append(pd.DataFrame(list(values), index=list(times), columns=[app]))
ax.plot(times, values, label=app)
if stack_xs is None:
stack_xs = times
stack_ys.append(values)
stack_labels.append(app)
ax.set_xlim(left=0, right=max_minutes)
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, bbox_to_anchor=(1.04, 1), loc="upper left")
output_name = output / f"{label}-{attr}-{region}.png"
fig.savefig(output_name.as_posix(), format='png', bbox_extra_artists=(lgd,), bbox_inches='tight')
plt.close(fig)
# write out CSV of the plot data
df = pd.concat(frames, axis=1)
df.to_csv(output / f"{label}-{attr}-{region}.csv", index_label="relative minutes")
if app_timestamps is not None and total_capacity is not None:
# create stacked plot
fig, ax = map_utils.subplots()
ax.set_title(f"{attr} {label} in {region}")
ax.set_xlabel("Time (minutes)")
ax.set_xlim(left=0, right=max_minutes)
ax.stackplot(stack_xs, stack_ys, labels=stack_labels)
total_capacity = map_utils.fill_missing_times(app_timestamps, total_capacity)
ax.plot(stack_xs, total_capacity, label="Total allocated capacity")
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, bbox_to_anchor=(1.04, 1), loc="upper left")
output_name = output / f"{label}-{attr}-{region}_stacked.png"
fig.savefig(output_name.as_posix(), format='png', bbox_extra_artists=(lgd,), bbox_inches='tight')
plt.close(fig)
elif 'QueueLength' != attr:
get_logger().warning("Missing data to draw stackplot for %s %s in %s", label, attr, region)
except:
get_logger().exception("Unexpected error")
def graph(output, first_timestamp, label, data_type, node_data):
"""
Graph network data.
Arguments:
output(Path): output directory
first_timestamp(int): first timestamp seen in the data
label(str): "Summary" or "Report"
label(str): "demand" or "load"
node_data(dict): service -> NetworkData
"""
for node_name, network_data in node_data.items():
rx_frames = list()
tx_frames = list()
all_frames = list()
rx_fig, rx_ax = map_utils.subplots()
rx_ax.set_title(f"{node_name} {label} {data_type} RX data")
rx_ax.set_xlabel("Time (minutes)")
tx_fig, tx_ax = map_utils.subplots()
tx_ax.set_title(f"{node_name} {label} {data_type} TX data")
tx_ax.set_xlabel("Time (minutes)")
all_fig, all_ax = map_utils.subplots()
all_ax.set_title(f"{node_name} {label} {data_type} ALL data")
all_ax.set_xlabel("Time (minutes)")
max_minutes = 0
rx_empty_plot = True
tx_empty_plot = True
all_empty_plot = True
for service, service_data in network_data.service.items():
if len(service_data.rx) > 0:
rx_empty_plot = False
rx_pairs = sorted(service_data.rx.items())
rx_timestamps, rx_values = zip(*rx_pairs)
rx_times = [
map_utils.timestamp_to_minutes(float(t) - first_timestamp)
for t in rx_timestamps
]
rx_series_label = f"{service} RX"
rx_ax.plot(rx_times, rx_values, label=rx_series_label)
rx_frames.append(
pd.DataFrame(list(rx_values),
index=rx_times,
columns=[rx_series_label]))
max_minutes = max(max_minutes, max(rx_times))
if len(service_data.tx) > 0:
tx_empty_plot = False
tx_pairs = sorted(service_data.tx.items())
tx_timestamps, tx_values = zip(*tx_pairs)
tx_times = [
map_utils.timestamp_to_minutes(float(t) - first_timestamp)
for t in tx_timestamps
]
tx_series_label = f"{service} TX"
tx_ax.plot(tx_times, tx_values, label=tx_series_label)
tx_frames.append(
pd.DataFrame(list(tx_values),
index=tx_times,
columns=[tx_series_label]))
max_minutes = max(max_minutes, max(tx_times))
if len(service_data.all_traffic) > 0:
all_empty_plot = False
all_pairs = sorted(service_data.all_traffic.items())
all_timestamps, all_values = zip(*all_pairs)
all_times = [
map_utils.timestamp_to_minutes(float(t) - first_timestamp)
for t in all_timestamps
]
all_series_label = f"{service} ALL"
all_ax.plot(all_times, all_values, label=all_series_label)
all_frames.append(
pd.DataFrame(list(all_values),
index=all_times,
columns=[all_series_label]))
max_minutes = max(max_minutes, max(all_times))
if not rx_empty_plot:
rx_ax.set_xlim(left=0, right=max_minutes)
rx_handles, rx_labels = rx_ax.get_legend_handles_labels()
rx_lgd = rx_ax.legend(rx_handles,
rx_labels,
bbox_to_anchor=(1.04, 1),
loc="upper left")
rx_output_name = output / f"network-{node_name}-{label}-RX-{data_type}.png"
rx_fig.savefig(rx_output_name.as_posix(),
format='png',
bbox_extra_artists=(rx_lgd, ),
bbox_inches='tight')
rx_df = pd.concat(rx_frames, axis=1)
rx_df.to_csv(output /
f"network-{node_name}-{label}-RX-{data_type}.csv",
index_label="relative minutes")
if not tx_empty_plot:
tx_ax.set_xlim(left=0, right=max_minutes)
tx_handles, tx_labels = tx_ax.get_legend_handles_labels()
tx_lgd = tx_ax.legend(tx_handles,
tx_labels,
bbox_to_anchor=(1.04, 1),
loc="upper left")
tx_output_name = output / f"network-{node_name}-{label}-TX-{data_type}.png"
tx_fig.savefig(tx_output_name.as_posix(),
format='png',
bbox_extra_artists=(tx_lgd, ),
bbox_inches='tight')
tx_df = pd.concat(tx_frames, axis=1)
tx_df.to_csv(output /
f"network-{node_name}-{label}-TX-{data_type}.csv",
index_label="relative minutes")
if not all_empty_plot:
all_ax.set_xlim(left=0, right=max_minutes)
all_handles, all_labels = all_ax.get_legend_handles_labels()
all_lgd = all_ax.legend(all_handles,
all_labels,
bbox_to_anchor=(1.04, 1),
loc="upper left")
all_output_name = output / f"network-{node_name}-{label}-ALL-{data_type}.png"
all_fig.savefig(all_output_name.as_posix(),
format='png',
bbox_extra_artists=(all_lgd, ),
bbox_inches='tight')
all_df = pd.concat(all_frames, axis=1)
all_df.to_csv(output /
f"network-{node_name}-{label}-ALL-{data_type}.csv",
index_label="relative minutes")
plt.close(rx_fig)
plt.close(tx_fig)
plt.close(all_fig)
def process_node(output, first_timestamp_ms, all_services, node_dir, show_load_demand_attr):
minutes = list()
planned_total = list() # number of planned containers total
planned_services = dict() # service -> list(value)
actual_total_allocation = list() # number of containers
actual_services_allocation = dict() # service -> list(allocation value)
actual_total_load = list() # load
actual_services_load = dict() # service -> list(load value)
actual_total_demand = list() # demand
actual_services_demand = dict() # service -> list(demand value)
ncp = node_dir.stem
get_logger().info("Processing directory for node %s: %s", ncp, node_dir)
for time_dir in sorted(node_dir.iterdir()):
if not time_dir.is_dir():
continue
get_logger().debug("Working on %s", time_dir)
directory_time = int(time_dir.stem)
directory_time_min = map_utils.timestamp_to_minutes(directory_time - first_timestamp_ms)
rlg_plan_file = time_dir / 'loadBalancerPlan.json'
if not rlg_plan_file.exists():
continue
resource_reports_file = time_dir / 'regionResourceReports-SHORT.json'
if not resource_reports_file.exists():
continue
try:
with open(rlg_plan_file, 'r') as f:
rlg_plan = json.load(f)
except json.decoder.JSONDecodeError:
get_logger().warning("Problem reading %s, skipping", rlg_plan_file)
continue
try:
with open(resource_reports_file, 'r') as f:
resource_reports = json.load(f)
except json.decoder.JSONDecodeError:
get_logger().warning("Problem reading %s, skipping", resource_reports_file)
continue
minutes.append(directory_time_min)
(p_total, p_per_service) = process_rlg_plan(rlg_plan)
planned_total.append(p_total)
for service in all_services:
service_list = planned_services.get(service, list())
if service in p_per_service:
service_list.append(p_per_service[service])
else:
service_list.append(0)
planned_services[service] = service_list
(a_total_allocation, a_per_service_allocation, a_total_load, a_per_service_load, a_total_demand, a_per_service_demand) = process_reports(resource_reports, show_load_demand_attr)
actual_total_allocation.append(a_total_allocation)
actual_total_load.append(a_total_load)
actual_total_demand.append(a_total_demand)
for service in all_services:
service_list_allocation = actual_services_allocation.get(service, list())
if service in a_per_service_allocation:
service_list_allocation.append(a_per_service_allocation[service])
else:
service_list_allocation.append(0)
actual_services_allocation[service] = service_list_allocation
service_list_load = actual_services_load.get(service, list())
if service in a_per_service_load:
service_list_load.append(a_per_service_load[service])
else:
service_list_load.append(0)
actual_services_load[service] = service_list_load
service_list_demand = actual_services_demand.get(service, list())
if service in a_per_service_demand:
service_list_demand.append(a_per_service_demand[service])
else:
service_list_demand.append(0)
actual_services_demand[service] = service_list_demand
graph(output, ncp, 'total', minutes, planned_total, actual_total_allocation, actual_total_load, actual_total_demand, show_load_demand_attr)
for service in all_services:
graph(output, ncp, service, minutes, planned_services[service], actual_services_allocation[service], actual_services_load[service], actual_services_demand[service], show_load_demand_attr)